feat(catalog-console): a keys status command, and complete rotation without a red main (#62, #68)

Two problems from issue #68's follow-up review:

1. The maintainer -- the only person who will ever use this tool -- cannot
   reliably tell which of the two signing keys is which or what state
   either is in. He already pasted a private key into a chat window
   because a prompt was ambiguous. That's a defect in this tool, not user
   error.

2. PR #71 rotates bcc_core.CATALOG_PUBKEYS, which makes
   data/catalog.json.sig (signed by the retired key) stop verifying and
   the CI catalog-signature job go red. The Console could previously only
   load/sign against `main`, so the only way through was to merge a red
   PR and fix main afterwards -- normalizing exactly the alarm fatigue
   this whole design exists to prevent.

Task 1 -- `python catalog_console.py keys`:
  A plain-English-first status report for BOTH keys: purpose, where the
  private half lives, whether it exists locally, its fingerprint, whether
  that fingerprint matches every place its public half is expected to be
  committed (bcc_core.CATALOG_PUBKEYS, ci.yml's trust anchor, and
  scripts/sign_checksums.RELEASE_PUBKEYS -- checked independently, since
  issue #68 finding 4 was exactly bcc_core.py and ci.yml silently
  drifting apart), and whether data/catalog.json.sig currently verifies --
  ending with the exact command to run next. Needs no passphrase and never
  touches private key bytes: a plaintext public-key cache
  (store_public_key/load_public_key) is written alongside the existing
  encrypted private blob at keygen time, precisely so this command can
  report a fingerprint without decrypting anything.

  The status/report logic (key_status, render_key_status_report,
  recommend_next_steps, fingerprint_pubkey, extract_pubkey_list_literal,
  extract_ci_trust_anchor_pubkey) is pure and lives in catalog_review.py;
  cmd_keys in catalog_console.py is a thin printer over it, per the
  project's existing pure-core/thin-GUI split.

Task 2 -- rotation completable without a red main:
  ReviewWindow now offers a "current branch" source (auto-detected via
  `current_branch()`, or --ref to name one explicitly) alongside "main"
  and open PRs. Loading it runs the exact same diff-against-last-signed /
  rotation-detection logic "main" always used (_load_own_ref, extracted
  from the old hardcoded-to-main _on_load), just parameterized on the
  ref. Signing now pushes to session.loaded_ref, never a hardcoded "main"
  (commit_and_push_signed_catalog's branch param was already there --
  only the call site was wrong). The ref-list computation itself is a
  pure function (compute_own_refs) so this seam is unit-testable without
  git or Qt. None of can_sign()'s guards (empty-diff, acknowledge-all,
  blocking-risk, TOCTOU) were touched.

  This lets a rotation branch be reviewed, re-attested (every entry,
  since the new key never vouched for any of them -- issue #68 finding 5
  follow-up), signed, and pushed to ITS OWN branch before it's ever
  merged.

Task 3 -- label the keys everywhere:
  PassphraseDialog now shows which key (CATALOG vs RELEASE) and its
  fingerprint before the passphrase field, both in its window title and
  its prompt text -- the exact ambiguity that led to a private key being
  pasted into a chat window. cmd_keygen's stored-key confirmation now
  reads "CATALOG private key encrypted..." / "RELEASE private key
  encrypted..." instead of a capitalized-lowercase kind. The reattest
  banner now says "CATALOG signing key" / "CATALOG key" throughout
  instead of "the key".

PySide6's import is now guarded (try/except -> _PYSIDE6_AVAILABLE) and
every GUI class definition that depends on it moved under
`if _PYSIDE6_AVAILABLE:`. `keygen`, `show-seed-b64`, and the new `keys`
command have no GUI dependency and now work (and are testable) in an
environment without PySide6 -- which is exactly this repo's own `test`
CI job (pytest + cryptography only, no PySide6). `gui` fails with a clear
message instead of an ImportError stack trace if it's missing.

Tests: 26 new pure-function tests in tests/test_catalog_review.py
(fingerprint_pubkey, extract_pubkey_list_literal,
extract_ci_trust_anchor_pubkey, key_status, recommend_next_steps,
render_key_status_report) and a new tests/test_catalog_console_git.py
(14 tests) covering compute_own_refs, current_branch,
commit_and_push_signed_catalog's branch targeting, and
catalog_sig_status_on_disk against real local git repos -- importing
catalog_console.py directly, proving it works without PySide6. 400
passed, 1 skipped (pre-existing). ruff check / ruff format --check clean.

.github/workflows/ci.yml

@@ -62,8 +62,115 @@ jobs:
 
       # bcc_core has no GUI imports, so the test suite needs no PySide6 —
       # keeps CI fast and avoids Qt system-library headaches on the runner.
+      # cryptography is for tests/test_checksums.py (release signing helper).
       - name: Install test dependencies
-        run: pip install pytest
+        run: pip install pytest cryptography
 
       - name: Run tests
         run: python -m pytest -v
+
+  # ── Catalog signature gate (#61) ─────────────────────────────────────────
+  #
+  # data/catalog.json is a list of command+args entries that BCC writes into
+  # the user's Claude config, which Claude then EXECUTES. The catalog is only
+  # trusted if it carries a valid Ed25519 signature from the maintainer key.
+  #
+  # The threat this gate exists for is NOT an outsider pushing to the repo —
+  # it is the maintainer merging a friendly-looking PR without really reading
+  # it. A contributor can change catalog.json but cannot produce a matching
+  # signature, so a blindly-merged PR lands here as a RED BUILD within a
+  # minute, instead of quietly riding into the next release.
+  #
+  # Public-key verification only. No secret is used or needed.
+  catalog-signature:
+    name: Catalog signature
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install dependencies
+        run: pip install cryptography
+
+      # 🔴 TRUST ANCHOR — issue #68 finding 4.
+      #
+      # This step used to do `import bcc_core as c` FROM THE CHECKED-OUT PR
+      # BRANCH and verify the catalog against c.CATALOG_PUBKEYS — i.e. it
+      # trusted the public key shipped in the very diff it was reviewing. A
+      # PR that changed data/catalog.json AND bcc_core.CATALOG_PUBKEYS (to
+      # an attacker key, with a matching signature produced by the attacker's
+      # matching private key) went green, because there was nothing outside
+      # the PR's own content to check the key against. The gate's whole
+      # point is catching a friendly-looking PR the maintainer merges
+      # without really reading it — and that hole made it a two-file diff.
+      #
+      # EXPECTED_CATALOG_PUBKEY_B64 below is hardcoded HERE, in the workflow
+      # file, independent of whatever bcc_core.py says on the PR branch. It
+      # is intentionally the only line in this step that matters for
+      # security review: changing it changes what this gate is willing to
+      # trust. THIS CONSTANT IS A TRUST ANCHOR. A PR that changes this line
+      # in the same diff as a catalog change is exactly the attack this gate
+      # exists to prevent — review a change to this line on its own,
+      # never bundled with a catalog update.
+      #
+      # NOTE for the next key rotation: update EXPECTED_CATALOG_PUBKEY_B64
+      # below to the new key's base64 form, as its own reviewed change.
+      - name: Verify data/catalog.json.sig
+        env:
+          EXPECTED_CATALOG_PUBKEY_B64: "0s24PmkZcTT5yxNDdyTPHl5fyxArrHNPJKBjnXoQd8k="
+        run: |
+          python - <<'PY'
+          import base64, os, pathlib, sys
+          import bcc_core as c
+
+          expected_pubkey_b64 = os.environ["EXPECTED_CATALOG_PUBKEY_B64"]
+
+          raw = pathlib.Path("data/catalog.json").read_bytes()
+          sig_path = pathlib.Path("data/catalog.json.sig")
+
+          if not sig_path.exists():
+              sys.exit("FAIL: data/catalog.json.sig is missing. The catalog must be "
+                       "signed via the Catalog Console (#62) before it can land.")
+
+          if b"\x00" * 32 in c.CATALOG_PUBKEYS:
+              sys.exit("FAIL: CATALOG_PUBKEYS still holds the placeholder key.")
+
+          # Trust anchor check FIRST, before verifying anything against
+          # bcc_core.CATALOG_PUBKEYS: a PR is not allowed to bring its own
+          # key. CATALOG_PUBKEYS on the checked-out branch must be EXACTLY
+          # the key(s) this workflow file itself expects -- no more, no
+          # fewer, no substitutions.
+          actual_pubkeys_b64 = [base64.b64encode(k).decode() for k in c.CATALOG_PUBKEYS]
+          if actual_pubkeys_b64 != [expected_pubkey_b64]:
+              sys.exit(
+                  "FAIL: bcc_core.CATALOG_PUBKEYS on this branch does not match the "
+                  "trust anchor hardcoded in .github/workflows/ci.yml.\n"
+                  f"  expected: {[expected_pubkey_b64]}\n"
+                  f"  actual:   {actual_pubkeys_b64}\n"
+                  "\n"
+                  "This PR is changing (or has changed) the catalog signing key. That "
+                  "change must be reviewed on its own, separately from any catalog "
+                  "content change, and the workflow's EXPECTED_CATALOG_PUBKEY_B64 "
+                  "updated deliberately -- not accepted because it happened to match "
+                  "whatever bcc_core.py says on this branch."
+              )
+
+          if not c.verify_catalog_signature(raw, sig_path.read_bytes(), c.CATALOG_PUBKEYS):
+              sys.exit(
+                  "FAIL: data/catalog.json does NOT match its signature.\n"
+                  "\n"
+                  "The catalog changed without being re-signed. Either someone edited\n"
+                  "it directly (a PR you merged?), or a signing pass was forgotten.\n"
+                  "Re-review and re-sign with the Catalog Console — do not bypass this."
+              )
+
+          problems = c.validate_catalog(c.load_catalog(raw))
+          if problems:
+              sys.exit("FAIL: catalog failed validation:\n  " + "\n  ".join(problems))
+
+          print("OK: catalog signature verifies, the pubkey matches the CI trust anchor, "
+                "and the catalog validates clean.")
+          PY

.github/workflows/release.yml

@@ -95,6 +95,92 @@ jobs:
           name: ${{ matrix.artifact }}
           path: ${{ matrix.artifact }}
 
+  # ── Signing-key smoke test (workflow_dispatch only) ─────────────────────
+  #
+  # The Publish job is gated on a tag, so a manual run never exercises the
+  # signing step — which means a wrong/missing RELEASE_SIGNING_KEY secret
+  # would only be discovered at the worst possible moment: during a real
+  # release. This job signs a throwaway manifest with the secret and verifies
+  # the result against scripts/sign_checksums.RELEASE_PUBKEYS.
+  #
+  # IMPORTANT (issue #68 finding 5): this must verify against the RELEASE
+  # public key, never bcc_core.CATALOG_PUBKEYS. The catalog key is the
+  # offline, maintainer-held root of trust for what BCC executes; it must
+  # NEVER be compared against a value that lives in a CI secret, because
+  # that comparison is itself a way to smuggle a catalog-trusted key through
+  # CI review ("does this repo secret match the catalog key" is a question
+  # this workflow must never even ask). The release key is a SEPARATE
+  # keypair, generated via `catalog_console.py keygen --release`, that only
+  # ever signs release SHA256SUMS manifests -- a CI/secret compromise burns
+  # this key, not the catalog key.
+  #
+  # It proves the two halves of the RELEASE keypair actually match, without
+  # publishing anything. Run it from the Actions tab after setting or
+  # rotating the secret.
+  signing-smoke-test:
+    name: Signing key smoke test
+    runs-on: ubuntu-latest
+    if: github.event_name == 'workflow_dispatch'
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      - name: Install dependencies
+        run: pip install cryptography
+
+      - name: Sign a throwaway manifest and verify against the RELEASE pubkey
+        env:
+          RELEASE_SIGNING_KEY: ${{ secrets.RELEASE_SIGNING_KEY }}
+        run: |
+          if [ -z "$RELEASE_SIGNING_KEY" ]; then
+            echo "FAIL: RELEASE_SIGNING_KEY secret is not set."
+            echo "Generate the RELEASE key (NOT the catalog key) with:"
+            echo "    python catalog_console.py keygen --release"
+            echo "then add its seed under Settings -> Actions -> Secrets, via:"
+            echo "    python catalog_console.py show-seed-b64 --release"
+            exit 1
+          fi
+          mkdir -p smoke && echo "smoke test payload" > smoke/hello.txt
+          python3 scripts/sign_checksums.py generate smoke --out smoke/SHA256SUMS
+          python3 scripts/sign_checksums.py sign --sums smoke/SHA256SUMS --out smoke/SHA256SUMS.sig
+          python - <<'PY'
+          import pathlib, sys
+          from scripts.sign_checksums import RELEASE_PUBKEYS, verify_checksums_against_any
+
+          # Deliberately does NOT import bcc_core / CATALOG_PUBKEYS at all --
+          # this smoke test must never be able to compare the CI secret
+          # against the catalog's root of trust (issue #68 finding 5). Only
+          # RELEASE_PUBKEYS (scripts/sign_checksums.py) is a legitimate
+          # target for a CI-resident key.
+          if not RELEASE_PUBKEYS:
+              sys.exit(
+                  "FAIL: scripts/sign_checksums.RELEASE_PUBKEYS is empty.\n"
+                  "\n"
+                  "Generate the release keypair with:\n"
+                  "    python catalog_console.py keygen --release\n"
+                  "then paste the printed public key into RELEASE_PUBKEYS in\n"
+                  "scripts/sign_checksums.py and commit that change."
+              )
+
+          sums = pathlib.Path("smoke/SHA256SUMS").read_text()
+          sig = pathlib.Path("smoke/SHA256SUMS.sig").read_bytes()
+
+          if not verify_checksums_against_any(RELEASE_PUBKEYS, sums, sig):
+              sys.exit(
+                  "FAIL: the signature produced by RELEASE_SIGNING_KEY does NOT verify\n"
+                  "against any key in scripts/sign_checksums.RELEASE_PUBKEYS.\n"
+                  "\n"
+                  "The secret and the shipped release public key are different keypairs.\n"
+                  "Downloaders would reject every signature this CI produces. Re-copy the\n"
+                  "seed from `catalog_console.py show-seed-b64 --release`, or update\n"
+                  "RELEASE_PUBKEYS with the matching public key."
+              )
+          print("OK: RELEASE_SIGNING_KEY matches a key in RELEASE_PUBKEYS.")
+          PY
+
   # ── Create GitHub Release with all three artifacts ──────────────────────
 
   release:
@@ -107,11 +193,76 @@ jobs:
       contents: write
 
     steps:
+      # Needed for scripts/sign_checksums.py — the release job otherwise
+      # only downloads build artifacts, it doesn't check out the repo.
+      - name: Checkout
+        uses: actions/checkout@v4
+
       - name: Download all artifacts
         uses: actions/download-artifact@v3
         with:
           path: artifacts
 
+      - name: Set up Python 3.12
+        uses: actions/setup-python@v5
+        with:
+          python-version: "3.12"
+
+      # download-artifact@v3 nests each artifact under a directory named
+      # after it (artifacts/<name>/<name>). Flatten into one directory so
+      # SHA256SUMS lists plain filenames, matching what `sha256sum -c`
+      # expects when run from inside an extracted release download.
+      - name: Collect release files
+        run: |
+          mkdir -p release-files
+          find artifacts -type f -exec cp {} release-files/ \;
+          ls -la release-files
+
+      - name: Generate SHA256SUMS
+        run: python3 scripts/sign_checksums.py generate release-files --out release-files/SHA256SUMS
+
+      # ── Sign the checksum manifest (best-effort) ──────────────────────
+      #
+      # BCC binaries are not code-signed (no budget for a paid cert). This
+      # is the free half: a checksum manifest, detached-signed with
+      # Ed25519, so a tampered download is detectable by anyone who
+      # checks. It does NOT remove Gatekeeper/SmartScreen warnings.
+      #
+      # The private key is a repo secret (RELEASE_SIGNING_KEY, base64 raw
+      # Ed25519 seed) for the RELEASE key -- a SEPARATE keypair from the
+      # catalog key, generated via `python catalog_console.py keygen
+      # --release` (issue #68 finding 5; #62). This key is intentionally
+      # CI-resident and signs ONLY this checksum manifest; it is never
+      # trusted to sign data/catalog.json. If it's not set, we still
+      # publish the release — just without a .sig — rather than fail the
+      # release outright.
+      - name: Check for signing key
+        id: signing
+        run: |
+          if [ -n "${{ secrets.RELEASE_SIGNING_KEY }}" ]; then
+            echo "has_key=true" >> "$GITHUB_OUTPUT"
+          else
+            echo "has_key=false" >> "$GITHUB_OUTPUT"
+          fi
+
+      - name: Install signing dependencies
+        if: steps.signing.outputs.has_key == 'true'
+        run: pip install cryptography
+
+      - name: Sign SHA256SUMS
+        if: steps.signing.outputs.has_key == 'true'
+        env:
+          RELEASE_SIGNING_KEY: ${{ secrets.RELEASE_SIGNING_KEY }}
+        run: |
+          python3 scripts/sign_checksums.py sign \
+            --sums release-files/SHA256SUMS \
+            --out release-files/SHA256SUMS.sig
+
+      - name: Warn — release will be unsigned
+        if: steps.signing.outputs.has_key != 'true'
+        run: |
+          echo "::warning::RELEASE_SIGNING_KEY secret is not set — this release is being published WITHOUT a signed SHA256SUMS.sig. Generate the RELEASE key (python catalog_console.py keygen --release) and add its seed (python catalog_console.py show-seed-b64 --release) as this secret before the next tag."
+
       - name: Create GitHub Release
         uses: softprops/action-gh-release@v2
         with:
@@ -119,7 +270,9 @@ jobs:
           draft: false
           prerelease: false
           generate_release_notes: false
-          files: artifacts/**/*
+          files: |
+            artifacts/**/*
+            release-files/SHA256SUMS*
           body: |
             ## Better Claude Config ${{ github.ref_name }}
 
@@ -139,5 +292,8 @@ jobs:
             xattr -cr /Applications/BetterClaudeConfig.app
             ```
 
+            ### Verifying your download
+            Every release includes `SHA256SUMS` (and, when the signing key is configured, a detached `SHA256SUMS.sig`). See [Verifying your download](https://git.avezzano.io/the_og/better-claude-config#verifying-your-download) in the README for commands. This proves you got the file we published — it does not remove Gatekeeper/SmartScreen warnings.
+
             ### Requirements
             No Python installation needed — the app is self-contained.

README.md

@@ -19,6 +19,135 @@ Pre-built self-contained binaries are attached to every [GitHub Release](../../r
 
 > **macOS Gatekeeper note:** the app is not notarized. On first launch, right-click → **Open**, or run `xattr -cr /Applications/BetterClaudeConfig.app` in a terminal.
 
+## Verifying your download
+
+BCC isn't code-signed — there's no budget for a paid certificate (macOS
+Developer ID, Windows Authenticode). Instead, every release publishes a
+`SHA256SUMS` file listing the checksum of each archive, detached-signed with
+Ed25519 as `SHA256SUMS.sig`. Both are attached to the release alongside the
+binaries.
+
+**What this proves:** the file you downloaded is byte-for-byte what we
+published, and the manifest itself was signed by our release key.
+
+**What this does NOT do:** it does not make the binary "safe," and it does
+**not** remove the macOS Gatekeeper or Windows SmartScreen warning — those
+are only suppressed by a paid OS-vendor certificate, which this project
+doesn't have. Verifying checksums is about detecting tampering in transit or
+on a mirror, not about vouching for the software.
+
+This manifest is signed with BCC's **release key**, which is a different
+key from the one that signs the MCP server catalog — see
+[Signing keys](#signing-keys) below for why, and for the public key value
+to use with `--pubkey-b64` below.
+
+### macOS / Linux
+
+```bash
+# From inside the folder you downloaded the release files into:
+sha256sum -c SHA256SUMS
+```
+
+If your `sha256sum` complains about missing files, download `SHA256SUMS`
+into the same directory as the archive you downloaded — it lists every
+platform's archive, and only the one(s) present will be checked.
+
+To also verify the manifest's signature (optional, requires Python +
+`pip install cryptography` and a checkout of this repo):
+
+```bash
+python3 scripts/sign_checksums.py verify \
+  --sums SHA256SUMS --sig SHA256SUMS.sig \
+  --pubkey-b64 "<the release public key from Signing keys, below>"
+```
+
+### Windows (PowerShell)
+
+```powershell
+Get-FileHash .\BetterClaudeConfig-Windows.zip -Algorithm SHA256
+```
+
+Compare the printed hash (case-insensitively) against the matching line in
+`SHA256SUMS`.
+
+### If a release has no `SHA256SUMS.sig`
+
+The signing key is a repo secret that has to be configured manually; if a
+release is missing the `.sig` file, the checksums themselves are still
+valid and safe to check against — the release workflow only skips signing,
+never checksum generation.
+
+## Signing keys
+
+BCC uses **two separate Ed25519 keypairs**, deliberately never the same
+key, because they protect different things and live in different places:
+
+| | Catalog key | Release key |
+|---|---|---|
+| Signs | `data/catalog.json` (the MCP server catalog every user's app trusts) | `SHA256SUMS` (the checksum manifest for release binaries) |
+| Verified by | `bcc_core.CATALOG_PUBKEYS` | `scripts/sign_checksums.RELEASE_PUBKEYS` |
+| Lives | Offline, passphrase-encrypted, maintainer's machine only (OS keychain or an encrypted file outside the repo — see the [Catalog Console](#files), issue #62) | A Gitea Actions repo secret, `RELEASE_SIGNING_KEY` — **intentionally CI-resident** |
+| Generated with | `python catalog_console.py keygen` | `python catalog_console.py keygen --release` |
+| Exported for CI with | *(never — there is no supported way to export this key)* | `python catalog_console.py show-seed-b64 --release` |
+
+**Confused about which key is which, or what state either is in?** Run:
+
+```bash
+python catalog_console.py keys
+```
+
+It needs no passphrase (it never touches private key bytes) and prints a
+plain-English report for both keys: where each private half lives, whether
+it's present on this machine, its fingerprint, whether that fingerprint
+matches what's actually committed in `bcc_core.py`, `ci.yml`'s trust
+anchor, and `scripts/sign_checksums.py`, and whether
+`data/catalog.json.sig` currently verifies — ending with the exact command
+to run next for whatever state it finds. This is the check that would have
+caught [issue #68](../../issues/68)'s finding 5 incident before it happened.
+
+**Why two keys:** the catalog key is the root of trust for what BCC
+actually *executes* on a user's machine — every `command`/`args` pair in
+the shipped catalog is only there because this key signed it. If that key
+and the release-checksum key were the same (as they briefly were — see
+[issue #68](../../issues/68)), then anything that can exfiltrate a Gitea
+Actions secret (a malicious workflow-file PR, a compromised runner, a leaky
+log) could sign a catalog every user's copy of BCC would trust, not just a
+checksum manifest. Splitting them means **a CI/secret compromise burns the
+release key, never the catalog key** — checksums for a future release could
+be forged, which is bad, but no attacker gains the ability to make BCC run
+arbitrary commands on installs that trust the catalog. That asymmetry is
+the entire point of having two keys instead of one.
+
+The catalog key is **never** meant to leave the maintainer's machine: it's
+generated, stored, unlocked, and used to sign entirely inside the Catalog
+Console (`catalog_console.py`), and `catalog_console.py show-seed-b64`
+refuses to run without `--release` specifically so the catalog seed can't
+be exported by habit or muscle memory.
+
+**Release signing public key** (Ed25519, base64, raw 32 bytes) — this is
+the RELEASE key, which signs `SHA256SUMS` (release checksums). It does
+**not** sign `data/catalog.json` and is not the key `bcc_core.CATALOG_PUBKEYS`
+trusts:
+
+```
+6BnPgJEHJFyVltFoLTCNadIsehjy00iiW8IRlC1TfhA=
+```
+
+The catalog public key (Ed25519, base64, raw 32 bytes) — this is the key
+that signs `data/catalog.json` and is trusted via `bcc_core.CATALOG_PUBKEYS`
+and the CI trust anchor in `.github/workflows/ci.yml`. It is listed here
+for completeness, not because you need it to verify a download — use the
+*release* key above for that:
+
+```
+0s24PmkZcTT5yxNDdyTPHl5fyxArrHNPJKBjnXoQd8k=
+```
+
+Both keys above were rotated 2026-07 — see [issue #68](../../issues/68)
+finding 5. The prior (shared) key is retired and is deliberately **not**
+kept in either trust list; retaining a burned key would defeat the point
+of rotating it.
+
 ## Run from source
 
 ```bash
@@ -92,6 +221,8 @@ file is also listed, marked *legacy*, so you can copy them over.
 - `test_core.py` — unit suite for the core (`python test_core.py`).
 - `bcc.spec` — PyInstaller build spec (cross-platform).
 - `scripts/build_icons.py` — regenerates `icons/app.icns` and `icons/app.ico` from source PNGs.
+- `scripts/sign_checksums.py` — generates and Ed25519-signs the release `SHA256SUMS` manifest (see [Verifying your download](#verifying-your-download)).
+- `catalog_console.py` / `catalog_review.py` — **maintainer-only**, never shipped to users (excluded from `bcc.spec`; see `tests/test_catalog_console_packaging.py`). The Catalog Console: review + sign `data/catalog.json` (against `main`, an open PR, or the branch you have checked out — `--ref <branch>` to be explicit, e.g. mid key-rotation, so a rotation can be signed and pushed to its own branch *before* it's merged, never forcing a red `main`), generate/manage both signing keys (`keygen`, `keygen --release`), and report on their status (`keys`, no passphrase needed) — see [Signing keys](#signing-keys).
 
 ## Building from source
 

bcc.py

@@ -23,6 +23,7 @@ from PySide6.QtGui import (
     QDesktopServices,
     QGuiApplication,
     QIcon,
+    QKeySequence,
     QPainter,
     QPixmap,
 )
@@ -38,6 +39,7 @@ from PySide6.QtWidgets import (
     QGridLayout,
     QHBoxLayout,
     QHeaderView,
+    QInputDialog,
     QLabel,
     QLineEdit,
     QListWidget,
@@ -1686,11 +1688,45 @@ class MainWindow(QMainWindow):
         head.setObjectName("h1")
         v.addWidget(head)
 
+        search_row = QHBoxLayout()
         self.search_box = QLineEdit()
         self.search_box.setPlaceholderText("Search servers by name, command, or url…")
         self.search_box.setClearButtonEnabled(True)
         self.search_box.textChanged.connect(self._on_search_changed)
-        v.addWidget(self.search_box)
+        search_row.addWidget(self.search_box, 1)
+        self.enable_all_btn = QPushButton("All on")
+        self.enable_all_btn.setToolTip("Enable every server")
+        self.enable_all_btn.clicked.connect(lambda: self._set_all_enabled(True))
+        search_row.addWidget(self.enable_all_btn)
+        self.disable_all_btn = QPushButton("All off")
+        self.disable_all_btn.setToolTip("Disable every server")
+        self.disable_all_btn.clicked.connect(lambda: self._set_all_enabled(False))
+        search_row.addWidget(self.disable_all_btn)
+        v.addLayout(search_row)
+
+        # Named server sets (issue #52): apply a saved Active/Disabled split
+        # in one click. Sets live in the config file under _bccServerSets.
+        sets_row = QHBoxLayout()
+        sets_lbl = QLabel("Set")
+        sets_lbl.setObjectName("muted")
+        sets_row.addWidget(sets_lbl)
+        self.sets_combo = QComboBox()
+        self.sets_combo.setMinimumWidth(120)
+        sets_row.addWidget(self.sets_combo, 1)
+        self.apply_set_btn = QPushButton("Apply")
+        self.apply_set_btn.setToolTip("Enable exactly this set's servers; disable the rest")
+        self.apply_set_btn.clicked.connect(self._apply_selected_set)
+        sets_row.addWidget(self.apply_set_btn)
+        self.save_set_btn = QPushButton("Save set…")
+        self.save_set_btn.setToolTip("Save the current Active/Disabled split as a named set")
+        self.save_set_btn.clicked.connect(self._save_set)
+        sets_row.addWidget(self.save_set_btn)
+        self.del_set_btn = QPushButton("−")
+        self.del_set_btn.setToolTip("Delete the selected set")
+        self.del_set_btn.setMaximumWidth(32)
+        self.del_set_btn.clicked.connect(self._delete_set)
+        sets_row.addWidget(self.del_set_btn)
+        v.addLayout(sets_row)
 
         # Active and Disabled sections live in a vertical splitter so the user
         # can drag the divider instead of being stuck with a fixed-height
@@ -1774,6 +1810,12 @@ class MainWindow(QMainWindow):
         undo_action.setShortcut("Ctrl+Z")
         undo_action.triggered.connect(self._undo)
         self.addAction(undo_action)
+        # Ctrl+S / Cmd+S shortcut — routed through a guard so it respects
+        # the same dirty/validation gating as the Save button.
+        save_action = QAction(self)
+        save_action.setShortcut(QKeySequence.StandardKey.Save)
+        save_action.triggered.connect(self._save_shortcut)
+        self.addAction(save_action)
         bar.addStretch()
         self.validation_lbl = QLabel("")
         bar.addWidget(self.validation_lbl)
@@ -1800,6 +1842,19 @@ class MainWindow(QMainWindow):
         self._mark_dirty()
         self.status.setText("Undone.")
 
+    def _set_all_enabled(self, enabled: bool):
+        """Flip every server's enabled flag in one step (one undo snapshot)."""
+        if not self.servers or all(s.enabled == enabled for s in self.servers):
+            return  # nothing to change
+        cur = self._current_index()
+        self._push_undo()
+        for s in self.servers:
+            s.enabled = enabled
+        sel = cur if 0 <= cur < len(self.servers) else None
+        self._refresh_tables(select_index=sel)
+        self._mark_dirty()
+        self.status.setText("All servers enabled." if enabled else "All servers disabled.")
+
     # --- profiles -------------------------------------------------------- #
     def reload_profiles(self):
         discovered = core.discover_profiles()
@@ -1892,6 +1947,7 @@ class MainWindow(QMainWindow):
         self._undo_stack.clear()
         self.undo_btn.setEnabled(False)
         self._health.clear()  # health results are per-profile; a fresh load invalidates them
+        self._refresh_sets_combo()  # sets are per-config; repopulate from the loaded file
         self._refresh_tables(select_index=0 if self.servers else -1)
         self._update_status(saved=False)
         if repaired:
@@ -1991,6 +2047,73 @@ class MainWindow(QMainWindow):
         cur = self._current_index()
         self._refresh_tables(select_index=cur if cur >= 0 else None)
 
+    # --- named server sets (issue #52) ------------------------------------ #
+    def _refresh_sets_combo(self, select: str | None = None):
+        sets = core.list_server_sets(self.full_config)
+        self.sets_combo.blockSignals(True)
+        self.sets_combo.clear()
+        for name in sorted(sets):
+            self.sets_combo.addItem(name)
+        if select is not None:
+            idx = self.sets_combo.findText(select)
+            if idx >= 0:
+                self.sets_combo.setCurrentIndex(idx)
+        self.sets_combo.blockSignals(False)
+        has_sets = bool(sets)
+        self.apply_set_btn.setEnabled(has_sets)
+        self.del_set_btn.setEnabled(has_sets)
+
+    def _apply_selected_set(self):
+        name = self.sets_combo.currentText()
+        sets = core.list_server_sets(self.full_config)
+        if name not in sets:
+            return
+        self._push_undo()
+        missing = core.apply_server_set(self.servers, sets[name])
+        self._refresh_tables(select_index=self._current_index() if self.servers else None)
+        self._mark_dirty()
+        on = sum(1 for s in self.servers if s.enabled)
+        msg = f"Applied set “{name}”  ·  {on} enabled. Review and Save."
+        if missing:
+            msg += f"  ⚠ no longer in this config: {', '.join(missing)}"
+        self.status.setText(msg)
+
+    def _save_set(self):
+        name, ok = QInputDialog.getText(
+            self,
+            "Save server set",
+            "Set name (saves which servers are currently Active):",
+            text=self.sets_combo.currentText(),
+        )
+        name = name.strip()
+        if not ok or not name:
+            return
+        if name in core.list_server_sets(self.full_config) and (
+            QMessageBox.question(self, "Set exists", f"Replace set “{name}”?")
+            != QMessageBox.StandardButton.Yes
+        ):
+            return
+        members = core.save_server_set(self.full_config, name, self.servers)
+        self._refresh_sets_combo(select=name)
+        self._mark_dirty()  # the set is written on the next Save
+        self.status.setText(
+            f"Set “{name}” saved ({len(members)} server(s)). Press Save to write it."
+        )
+
+    def _delete_set(self):
+        name = self.sets_combo.currentText()
+        if not name:
+            return
+        if (
+            QMessageBox.question(self, "Delete set", f"Delete set “{name}”?")
+            != QMessageBox.StandardButton.Yes
+        ):
+            return
+        if core.delete_server_set(self.full_config, name):
+            self._refresh_sets_combo()
+            self._mark_dirty()
+            self.status.setText(f"Set “{name}” deleted. Press Save to write the change.")
+
     # --- test all (spawn-test every enabled local server) ---------------- #
     def _test_all_servers(self):
         targets = [s for s in self.servers if s.enabled and s.kind == "stdio"]
@@ -2309,11 +2432,23 @@ class MainWindow(QMainWindow):
             self.validation_lbl.setStyleSheet(f"color: {WARN};")
             self.save_btn.setEnabled(False)
             return False
-        self.validation_lbl.setText("✓ valid")
+        lint_warnings = core.lint_servers(self.servers)
-        self.validation_lbl.setStyleSheet(f"color: {GOOD};")
+        if lint_warnings:
+            self.validation_lbl.setText(f"⚠ {lint_warnings[0]}")
+            self.validation_lbl.setStyleSheet(f"color: {WARN};")
+        else:
+            self.validation_lbl.setText("✓ valid")
+            self.validation_lbl.setStyleSheet(f"color: {GOOD};")
         self.save_btn.setEnabled(self.dirty)
         return True
 
+    def _save_shortcut(self):
+        """Ctrl+S / Cmd+S handler — only fires when the Save button itself
+        would accept a click, so the shortcut can't bypass validation/dirty
+        gating."""
+        if self.save_btn.isEnabled():
+            self.save()
+
     def save(self):
         if not self.current_profile:
             return

bcc.spec

@@ -31,7 +31,7 @@ a = Analysis(
     ["bcc.py"],
     pathex=[],
     binaries=[],
-    datas=[("icons", "icons")],
+    datas=[("icons", "icons"), ("data/catalog.json", "data")],
     hiddenimports=[],
     hookspath=[],
     hooksconfig={},
@@ -78,8 +78,8 @@ if sys.platform == "darwin":
         info_plist={
             "CFBundleName": "Better Claude Config",
             "CFBundleDisplayName": "Better Claude Config",
-            "CFBundleShortVersionString": "1.2.1",
+            "CFBundleShortVersionString": "1.3.0",
-            "CFBundleVersion": "1.2.1",
+            "CFBundleVersion": "1.3.0",
             "NSHighResolutionCapable": True,
             "NSRequiresAquaSystemAppearance": False,  # supports dark mode
             "LSMinimumSystemVersion": "11.0",

bcc_core.py

@@ -13,6 +13,7 @@ in its original position.
 
 from __future__ import annotations
 
+import base64
 import contextlib
 import difflib
 import functools
@@ -32,6 +33,9 @@ from pathlib import Path
 from typing import NamedTuple
 from urllib.parse import urlparse
 
+from cryptography.exceptions import InvalidSignature
+from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PublicKey
+
 CONFIG_FILENAME = "claude_desktop_config.json"
 
 # Disabled servers are parked under this non-standard key. Claude Desktop only
@@ -39,6 +43,12 @@ CONFIG_FILENAME = "claude_desktop_config.json"
 # we can toggle it back on without losing the definition.
 DISABLED_KEY = "_disabledMcpServers"
 
+# Named server sets: {set_name: [enabled server names]}. Same pattern as
+# DISABLED_KEY — a bcc-owned key Claude ignores, stored in the config file so
+# sets travel with it. Applying a set enables exactly the listed servers and
+# parks the rest under DISABLED_KEY.
+SETS_KEY = "_bccServerSets"
+
 BACKUP_DIRNAME = ".bcc_backups"
 MAX_BACKUPS = 15
 
@@ -59,7 +69,7 @@ KNOWN_FIELDS = {"command", "args", "env", "url", "type", "headers"}
 # binary. All network I/O here is fail-quiet (returns None on any problem)
 # so it's safe to run unattended, off the UI thread, at startup.
 # --------------------------------------------------------------------------- #
-__version__ = "1.2.1"
+__version__ = "1.3.0"
 
 REPO_URL = "https://git.avezzano.io/the_og/better-claude-config"
 ISSUES_URL = f"{REPO_URL}/issues"
@@ -282,10 +292,43 @@ def msix_warning_text(
     )
 
 
+def discover_project_configs(claude_json_path: str | os.PathLike) -> list[Profile]:
+    """
+    Find project-scope `.mcp.json` configs known to Claude Code.
+
+    `~/.claude.json` keeps a `projects` map keyed by absolute project
+    directory path (that's what the CLI writes as it's used in each repo).
+    Any project whose directory has a `.mcp.json` file next to it -- a
+    standalone file with a top-level `mcpServers` object, same shape BCC
+    already edits -- is surfaced here as its own profile so it can be opened
+    via 'Add config...' without hunting for the path by hand.
+
+    Fails quiet: a missing/unreadable/malformed `claude_json_path`, or a
+    `projects` value that isn't a dict, just yields an empty list rather than
+    raising -- this is best-effort discovery, not a required config load.
+    """
+    try:
+        cfg = load_config(claude_json_path)
+    except Exception:
+        return []
+    projects = cfg.get("projects")
+    if not isinstance(projects, dict):
+        return []
+    out: list[Profile] = []
+    for key in sorted(k for k in projects if isinstance(k, str)):
+        mcp_path = Path(key) / ".mcp.json"
+        if mcp_path.is_file():
+            out.append(
+                Profile(label=f"Project: {Path(key).name}", path=mcp_path, config_exists=True)
+            )
+    return out
+
+
 def discover_profiles() -> list[Profile]:
     """
     Find every `Claude*` data directory in the platform's app-support base
-    (Claude Desktop installs), then also check for a Claude Code global config.
+    (Claude Desktop installs), then also check for a Claude Code global config
+    and any project-scope `.mcp.json` configs it knows about.
 
     Claude Desktop: scans the platform app-support folder for any `Claude*`
     directory (catches `Claude`, `Claude-Work`, etc.).
@@ -298,6 +341,10 @@ def discover_profiles() -> list[Profile]:
     `claude mcp add` writes; project scope is a per-repo .mcp.json, which can
     be opened via 'Add config…'). NOT ~/.claude/settings.json — that file is
     for permissions/hooks and rejects an mcpServers key with a schema error.
+    Project scope: ~/.claude.json also tracks a `projects` map, one entry per
+    directory Claude Code has been run in; any of those with a `.mcp.json`
+    file are surfaced as their own profiles too (see
+    `discover_project_configs`).
     """
     base = app_support_base()
     out: list[Profile] = []
@@ -319,6 +366,12 @@ def discover_profiles() -> list[Profile]:
     cc_cfg = home / ".claude.json"
     out.append(Profile(label="Claude Code", path=cc_cfg, config_exists=cc_cfg.is_file()))
 
+    existing_paths = {str(p.path) for p in out}
+    for proj in discover_project_configs(cc_cfg):
+        if str(proj.path) not in existing_paths:
+            existing_paths.add(str(proj.path))
+            out.append(proj)
+
     # Legacy: earlier BCC versions (and hand-edits) may have parked servers in
     # ~/.claude/settings.json, where Claude Code ignores them. Surface that
     # file only when it actually contains an mcpServers block, so the user can
@@ -409,6 +462,70 @@ def extract_servers(cfg: dict) -> list[ServerEntry]:
     return out
 
 
+# --------------------------------------------------------------------------- #
+# Named server sets (issue #52)
+# --------------------------------------------------------------------------- #
+def list_server_sets(cfg: dict) -> dict[str, list[str]]:
+    """
+    Return {set_name: [enabled server names]} from cfg's SETS_KEY.
+
+    Fail-soft: entries whose value isn't a list of strings (hand-edited or
+    corrupted) are skipped rather than raising, so one bad set never hides
+    the rest.
+    """
+    raw = cfg.get(SETS_KEY)
+    if not isinstance(raw, dict):
+        return {}
+    out: dict[str, list[str]] = {}
+    for name, members in raw.items():
+        if isinstance(members, list) and all(isinstance(m, str) for m in members):
+            out[str(name)] = list(members)
+    return out
+
+
+def save_server_set(cfg: dict, name: str, servers: list[ServerEntry]) -> list[str]:
+    """
+    Snapshot the current enabled-server names into cfg under SETS_KEY as
+    `name` (overwriting an existing set of that name). Returns the saved
+    member list. The caller decides when cfg reaches disk (normal Save flow).
+    """
+    members = [s.name for s in servers if s.enabled]
+    sets = cfg.get(SETS_KEY)
+    if not isinstance(sets, dict):
+        sets = {}
+        cfg[SETS_KEY] = sets
+    sets[name] = members
+    return members
+
+
+def delete_server_set(cfg: dict, name: str) -> bool:
+    """Remove set `name` from cfg. Drops SETS_KEY entirely when the last set
+    goes, so untouched configs don't grow an empty bcc key. Returns True if
+    something was deleted."""
+    sets = cfg.get(SETS_KEY)
+    if not isinstance(sets, dict) or name not in sets:
+        return False
+    del sets[name]
+    if not sets:
+        cfg.pop(SETS_KEY, None)
+    return True
+
+
+def apply_server_set(servers: list[ServerEntry], enabled_names: list[str]) -> list[str]:
+    """
+    Enable exactly the servers named in `enabled_names`; disable every other
+    entry (in place). Returns the set members that no longer exist in
+    `servers` — the caller surfaces those as a warning, and the rest of the
+    set still applies.
+    """
+    wanted = set(enabled_names)
+    present: set[str] = set()
+    for s in servers:
+        s.enabled = s.name in wanted
+        present.add(s.name)
+    return sorted(wanted - present)
+
+
 def resolve_name_collision(name: str, existing: set[str]) -> str:
     """
     Return a name guaranteed not to collide with `existing`.
@@ -1213,6 +1330,63 @@ def validate_servers(servers: list[ServerEntry]) -> list[str]:
     return problems
 
 
+def lint_server(name: str, data: dict) -> list[str]:
+    """Return non-blocking structural warnings for a single server definition.
+
+    Unlike validate_servers, nothing here blocks Save -- these are advisory
+    notes about shapes that will round-trip through JSON fine but are
+    probably not what the user intended (args given as a plain string
+    instead of a list, an env value that isn't a string, an unrecognized
+    `type`, unknown top-level fields, etc.).
+    """
+    nm = name.strip() or "(unnamed)"
+    warnings: list[str] = []
+
+    if "command" in data and not isinstance(data["command"], str):
+        warnings.append(f"'{nm}': 'command' should be a string")
+
+    if "args" in data:
+        args = data["args"]
+        if not isinstance(args, list):
+            warnings.append(f"'{nm}': 'args' should be a list (one argument per item)")
+        elif any(not isinstance(a, str) for a in args):
+            warnings.append(
+                f"'{nm}': 'args' contains non-string values "
+                "(they will be saved as-is; Claude expects strings)"
+            )
+
+    for field in ("env", "headers"):
+        if field not in data:
+            continue
+        val = data[field]
+        if not isinstance(val, dict):
+            warnings.append(f"'{nm}': '{field}' should be an object of string key/value pairs")
+        elif any(not isinstance(v, str) for v in val.values()):
+            warnings.append(
+                f"'{nm}': '{field}' contains non-string values "
+                "(they will be saved as-is; Claude expects strings)"
+            )
+
+    if "type" in data:
+        t = data["type"]
+        if t not in ("http", "sse", "stdio"):
+            warnings.append(f"'{nm}': 'type' should be one of http, sse, stdio (found {t!r})")
+
+    extra = sorted(k for k in data if k not in KNOWN_FIELDS)
+    if extra:
+        warnings.append(f"'{nm}': extra fields preserved as-is: {', '.join(extra)}")
+
+    return warnings
+
+
+def lint_servers(servers: list[ServerEntry]) -> list[str]:
+    """Concatenate lint_server warnings across every entry, in order."""
+    out: list[str] = []
+    for s in servers:
+        out.extend(lint_server(s.name, s.data))
+    return out
+
+
 # --------------------------------------------------------------------------- #
 # Search / filter
 # --------------------------------------------------------------------------- #
@@ -1973,3 +2147,655 @@ def restart_claude_desktop() -> RestartResult:
     if sys.platform.startswith("win"):
         return _restart_claude_desktop_windows()
     return RestartResult(False, "Restarting Claude Desktop isn't supported on this platform.")
+
+
+# --------------------------------------------------------------------------- #
+# MCP server catalog (issue #10 / #61)
+#
+# A curated, SIGNED list of ready-to-use MCP server definitions (bundled with
+# the app and, later, fetchable/cacheable — see follow-up issues). Every
+# function here is pure and defensive: catalog bytes may come from a fetch
+# over the network, a disk cache, or the copy frozen into the binary, and
+# all three are treated as equally untrusted until their signature verifies.
+# --------------------------------------------------------------------------- #
+
+# Commands a catalog entry's config is allowed to launch. Anything else
+# (bash, sh, curl, a raw script interpreter that isn't on this list, ...)
+# is rejected by validate_catalog() regardless of how plausible it looks.
+CATALOG_ALLOWED_COMMANDS = frozenset({"npx", "uvx", "docker", "node", "python", "python3"})
+
+# Env var keys a catalog entry's config.env must never set. Every one of
+# these is a loader/interpreter override that lets a value walk straight
+# past CATALOG_ALLOWED_COMMANDS and the -e/--eval/-c deny-rule below: e.g.
+# NODE_OPTIONS="--require /tmp/x.js" turns an allowlisted `npx` entry into
+# arbitrary code execution without ever touching config.args, which is the
+# only field the allowlist/deny-rules/ASCII/secret checks used to cover.
+# Matched case-insensitively -- env keys are case-sensitive on POSIX, but a
+# `node_options` lookalike is exactly the kind of thing this must catch.
+CATALOG_DENIED_ENV_KEYS = frozenset(
+    {
+        "NODE_OPTIONS",
+        "PYTHONSTARTUP",
+        "PYTHONPATH",
+        "PYTHONHOME",
+        "LD_PRELOAD",
+        "LD_LIBRARY_PATH",
+        "DYLD_INSERT_LIBRARIES",
+        "DYLD_LIBRARY_PATH",
+        "BROWSER",
+        "PATH",
+        "NODE_REPL_EXTERNAL_MODULE",
+    }
+)
+
+# Ed25519 public keys allowed to sign a catalog, raw 32-byte form. A LIST
+# (not a single key) so keys can be rotated without bricking installs that
+# still trust an older key: verify_catalog_signature() accepts a match
+# against ANY key in this list.
+CATALOG_PUBKEYS: list[bytes] = [
+    base64.b64decode("0s24PmkZcTT5yxNDdyTPHl5fyxArrHNPJKBjnXoQd8k="),
+]
+
+# Domain-separation prefix for the signed message. The signature covers
+# this prefix + the raw catalog bytes, never the raw bytes alone, so a
+# catalog signature can't be replayed against some other byte-for-byte-
+# identical payload that means something else in a different context.
+_CATALOG_SIG_DOMAIN = b"bcc-catalog-v1|"
+
+# Top-level fields that must be https:// URLs when present.
+_CATALOG_URL_FIELDS = ("homepage", "docs_url", "source")
+
+# Inline secret-flag=value forms. Distinct from _TOKEN_PREFIXES below --
+# this catches "--api-key=<real value>" even when the value itself doesn't
+# match a well-known token prefix.
+_CATALOG_SECRET_ARG_RE = re.compile(r"(?i)--api[-_]?key=|--token=|--password=")
+
+# <PLACEHOLDER>-style tokens the GUI must have the user fill in before Save.
+_PLACEHOLDER_RE = re.compile(r"<[^<>\s]+>")
+
+# A catalog entry's id becomes an mcpServers JSON key AND is interpolated
+# into Qt.AutoText widgets (status bar, QMessageBox) -- an id like
+# "<b>Verified</b>" renders as markup there. Not RCE, but UI spoofing, so
+# ids are constrained to a plain lowercase slug.
+_CATALOG_ID_RE = re.compile(r"^[a-z0-9][a-z0-9._-]{0,63}$")
+
+# Docker flags that consume the next arg as a value (so that value must not
+# be mistaken for the image reference when locating it in config.args).
+_DOCKER_VALUE_FLAGS = frozenset(
+    {
+        "-e",
+        "--env",
+        "-v",
+        "--volume",
+        "-p",
+        "--publish",
+        "--name",
+        "-w",
+        "--workdir",
+        "-u",
+        "--user",
+        "--entrypoint",
+        "--network",
+        "--platform",
+        "--add-host",
+        "-l",
+        "--label",
+    }
+)
+
+# How many versions a single accepted catalog jump may leap in one go. Bounds
+# a "freeze" attack: a compromised/leaked signing key claiming an absurd
+# future version would otherwise permanently outrank every legitimate
+# catalog release from then on, since the resolver always prefers the
+# highest verified version.
+_CATALOG_MAX_VERSION_JUMP = 1000
+
+
+def load_catalog(raw: bytes | str) -> dict:
+    """
+    Parse catalog bytes/text into a dict using STRICT json.loads ONLY.
+
+    🔴 CRITICAL: the lenient JSON repair pipeline (repair_json_text,
+    parse_pasted_json / parse_pasted_json_verbose) must NEVER be wired in
+    here, or anywhere near catalog handling. That pipeline exists to be
+    forgiving of hand-pasted snippets from docs and blog posts — smart
+    quotes, trailing commas, unquoted keys, whatever a human fat-fingered.
+    Forgiveness is exactly the property a signed payload cannot have:
+    verify_catalog_signature() authenticates the exact bytes that were
+    signed. If what gets displayed/executed is a "repaired" reinterpretation
+    of those bytes rather than the bytes themselves, the signature check
+    still passes while guaranteeing nothing about what actually runs. Always
+    verify raw bytes, then load_catalog() those SAME raw bytes.
+    """
+    return json.loads(raw)
+
+
+def catalog_version(data: dict) -> int:
+    """Extract the integer version from a parsed catalog dict (0 if absent/bad)."""
+    version = data.get("version") if isinstance(data, dict) else None
+    return version if isinstance(version, int) and not isinstance(version, bool) else 0
+
+
+def _secret_looking_arg(a: str) -> bool:
+    """
+    True when a catalog arg string looks like it embeds a real secret. Reuses
+    the existing token-prefix detector (_is_secret_value / _TOKEN_PREFIXES)
+    rather than reimplementing it — one definition of "looks like a secret"
+    for the whole app.
+    """
+    if _CATALOG_SECRET_ARG_RE.search(a):
+        return True
+    value = a.split("=", 1)[1] if "=" in a else a
+    return _is_secret_value(value) or _is_secret_value(a)
+
+
+def _docker_arg_violations(tag: str, args: list[str]) -> list[str]:
+    """--privileged and volume mounts rooted at / or $HOME are refused."""
+    problems: list[str] = []
+    if "--privileged" in args:
+        problems.append(f"{tag}: config.args uses --privileged, which is not allowed.")
+
+    i = 0
+    while i < len(args):
+        a = args[i]
+        mount = None
+        if a in ("-v", "--volume") and i + 1 < len(args):
+            mount = args[i + 1]
+            i += 1
+        elif a.startswith("--volume="):
+            mount = a.split("=", 1)[1]
+        elif a.startswith("-v") and a != "-v":
+            mount = a[2:]
+        if mount:
+            source = mount.split(":", 1)[0]
+            if source in ("/", "$HOME") or source.startswith("$HOME"):
+                problems.append(
+                    f"{tag}: config.args mounts {source!r}, which is not allowed "
+                    "(volume mounts of / or $HOME are refused)."
+                )
+        i += 1
+    return problems
+
+
+def _catalog_package_spec_version(spec: str) -> str | None:
+    """
+    Extract the version pin from an npm-style package spec, or None if the
+    spec carries no pin.
+
+    Handles unscoped "name@version" and scoped "@scope/name@version" --
+    scoped names have a leading "@" that is NOT the version separator, so a
+    naive split on the first/only "@" misparses "@scope/pkg" (no version)
+    as pinned to "scope/pkg". Splitting from the right side instead is safe
+    for both forms because a package name may contain "@" only as the
+    scope's leading character.
+    """
+    if spec.startswith("@"):
+        rest = spec[1:]
+        if "@" not in rest:
+            return None
+        _, _, version = rest.rpartition("@")
+        return version or None
+    if "@" not in spec:
+        return None
+    _, _, version = spec.rpartition("@")
+    return version or None
+
+
+def _catalog_package_spec_pinned(spec: str) -> bool:
+    """
+    True if `spec` carries an exact version pin. Covers npm's "name@version"
+    / "@scope/name@version" and uv's documented PyPI pin forms
+    "name@version" and "name==version".
+    """
+    if "==" in spec:
+        _, _, version = spec.partition("==")
+        return bool(version)
+    return bool(_catalog_package_spec_version(spec))
+
+
+def _first_catalog_package_spec(args: list[str]) -> str | None:
+    """
+    The first arg that could plausibly BE a package spec: skip flags
+    (leading "-") and <PLACEHOLDER> tokens (which can't be validated and
+    are filled in by the user later, never shipped by the catalog as the
+    package name itself). Everything after the first hit is ignored --
+    trailing flags, paths, and placeholders are not package specs.
+    """
+    for a in args:
+        if a.startswith("-"):
+            continue
+        if _PLACEHOLDER_RE.fullmatch(a):
+            continue
+        return a
+    return None
+
+
+def _catalog_pin_violations(tag: str, command: str, args: list[str]) -> list[str]:
+    """
+    Version-pinning enforcement (finding #3): a catalog PR can otherwise
+    ship `npx -y @scope/pkg` or `docker run img:latest` and the *next*
+    resolve of that package/image is whatever the registry serves that day
+    -- outside review, outside the signature's meaning. This is the only
+    place that enforces pinning at runtime; catalog_review.py's
+    risk_unpinned_package() is a maintainer-facing hint, not a gate.
+    """
+    if command in ("npx", "uvx"):
+        spec = _first_catalog_package_spec(args)
+        if spec is None:
+            return [f"{tag}: config.args must include a package spec to pin (e.g. name@1.2.3)."]
+        if not _catalog_package_spec_pinned(spec):
+            return [
+                f"{tag}: config.args package {spec!r} is not version-pinned; use "
+                "name@version, @scope/name@version, or name==version."
+            ]
+        return []
+
+    if command == "docker":
+        image = _docker_image_ref(args)
+        if image is None:
+            return [f"{tag}: config.args docker command has no image reference to pin."]
+        _, sep, image_tag = image.rpartition(":")
+        if not sep or "/" in image_tag:
+            return [
+                f"{tag}: config.args docker image {image!r} has no explicit tag; "
+                "pin an exact version (not 'latest', not untagged)."
+            ]
+        if image_tag == "latest":
+            return [
+                f"{tag}: config.args docker image {image!r} uses the 'latest' tag, "
+                "which is not allowed; pin an exact version."
+            ]
+        return []
+
+    return []
+
+
+def _docker_image_ref(args: list[str]) -> str | None:
+    """
+    Locate the image reference in a `docker run ...` args list: skip the
+    "run" subcommand and any flags, including ones that consume the next
+    token as a value (-e, -v, --name, ...) so that value isn't mistaken for
+    the image. The first remaining positional token is the image.
+    """
+    i = 0
+    if i < len(args) and args[i] == "run":
+        i += 1
+    while i < len(args):
+        a = args[i]
+        if a.startswith("-"):
+            if a in _DOCKER_VALUE_FLAGS and "=" not in a:
+                i += 2
+            else:
+                i += 1
+            continue
+        return a
+    return None
+
+
+def _validate_catalog_config(tag: str, config) -> list[str]:
+    """Validate the `config` block of a basic-tier catalog entry."""
+    if not isinstance(config, dict):
+        return [f"{tag}: basic entries require a 'config' object with command+args."]
+
+    problems: list[str] = []
+
+    command = config.get("command")
+    if not isinstance(command, str) or not command:
+        problems.append(f"{tag}: config.command must be a non-empty string.")
+        command = ""
+    elif not command.isascii():
+        problems.append(f"{tag}: config.command must be ASCII (non-ASCII code points rejected).")
+
+    if command and command not in CATALOG_ALLOWED_COMMANDS:
+        problems.append(
+            f"{tag}: config.command {command!r} is not on the catalog allowlist "
+            f"({', '.join(sorted(CATALOG_ALLOWED_COMMANDS))})."
+        )
+
+    raw_args = config.get("args")
+    args_ok = isinstance(raw_args, list) and all(isinstance(a, str) for a in raw_args)
+    args = raw_args if args_ok else []
+    if not args_ok:
+        problems.append(f"{tag}: config.args must be a list of strings.")
+
+    for a in args:
+        if not a.isascii():
+            problems.append(f"{tag}: config.args contains a non-ASCII value ({a!r}).")
+        if _secret_looking_arg(a):
+            problems.append(
+                f"{tag}: config.args contains a secret-looking value ({a!r}); "
+                "secrets belong in env, never args."
+            )
+
+    if command in ("node", "python", "python3") and any(a in ("-e", "--eval", "-c") for a in args):
+        problems.append(
+            f"{tag}: config.args uses -e/--eval/-c with {command!r}, which is not allowed."
+        )
+
+    if command == "docker":
+        problems.extend(_docker_arg_violations(tag, args))
+
+    # Version pinning (finding #3) -- only meaningful once command/args are
+    # actually well-formed; a malformed args list already got its own
+    # problem above and has nothing left to pin-check.
+    if args_ok and command in ("npx", "uvx", "docker"):
+        problems.extend(_catalog_pin_violations(tag, command, args))
+
+    env = config.get("env")
+    if env is not None:
+        env_ok = isinstance(env, dict) and all(
+            isinstance(k, str) and isinstance(v, str) for k, v in env.items()
+        )
+        if not env_ok:
+            problems.append(f"{tag}: config.env must be an object of string values.")
+        else:
+            # config.env (finding #2): unlike args, env was previously
+            # type-checked ONLY -- no allowlist, no deny-rule, no ASCII
+            # check, no secret check. That made it the single easiest way
+            # to smuggle a payload past every other guard in this
+            # function: an allowlisted `command: npx` plus
+            # NODE_OPTIONS=--require /tmp/x.js in env walks straight past
+            # the command allowlist AND the -e/--eval/-c deny-rule above,
+            # because neither of those ever looks at env.
+            for key, value in env.items():
+                if not key.isascii():
+                    problems.append(
+                        f"{tag}: config.env key {key!r} must be ASCII "
+                        "(non-ASCII code points rejected)."
+                    )
+                if key.upper() in CATALOG_DENIED_ENV_KEYS:
+                    problems.append(
+                        f"{tag}: config.env key {key!r} is on the catalog deny-list "
+                        "(interpreter/loader override) and is not allowed."
+                    )
+                if not value.isascii():
+                    problems.append(
+                        f"{tag}: config.env value for {key!r} must be ASCII "
+                        "(non-ASCII code points rejected)."
+                    )
+                if _is_secret_value(value):
+                    problems.append(
+                        f"{tag}: config.env[{key!r}] looks like a real secret value; "
+                        "catalog entries must never ship secret values."
+                    )
+                if value != "" and not _PLACEHOLDER_RE.fullmatch(value):
+                    problems.append(
+                        f"{tag}: config.env[{key!r}] must be an empty string or a "
+                        "single <PLACEHOLDER> token -- the catalog declares which env "
+                        "vars a server needs, it never supplies their values."
+                    )
+
+    return problems
+
+
+def _validate_catalog_entry(idx: int, entry, seen_ids: set[str]) -> list[str]:
+    """Validate a single `servers[idx]` catalog entry."""
+    tag = f"servers[{idx}]"
+    if not isinstance(entry, dict):
+        return [f"{tag}: must be an object."]
+
+    problems: list[str] = []
+
+    entry_id = entry.get("id")
+    if not isinstance(entry_id, str) or not entry_id.strip():
+        problems.append(f"{tag}: 'id' must be a non-empty string.")
+    else:
+        tag = f"servers[{idx}] ({entry_id!r})"
+        if not entry_id.isascii():
+            problems.append(f"{tag}: 'id' must be ASCII (non-ASCII code points rejected).")
+        elif not _CATALOG_ID_RE.match(entry_id):
+            problems.append(
+                f"{tag}: 'id' must match ^[a-z0-9][a-z0-9._-]{{0,63}}$ "
+                "(it becomes an mcpServers JSON key and is interpolated into "
+                "Qt.AutoText widgets)."
+            )
+        if entry_id in seen_ids:
+            problems.append(f"{tag}: duplicate id.")
+        seen_ids.add(entry_id)
+
+    for field in ("display", "description", "category"):
+        if not isinstance(entry.get(field), str) or not entry[field].strip():
+            problems.append(f"{tag}: '{field}' must be a non-empty string.")
+
+    if not isinstance(entry.get("official"), bool):
+        problems.append(f"{tag}: 'official' must be a boolean.")
+
+    setup = entry.get("setup")
+    if setup not in ("basic", "link-only"):
+        problems.append(f"{tag}: 'setup' must be 'basic' or 'link-only'.")
+
+    env_required = entry.get("env_required")
+    if not isinstance(env_required, dict):
+        problems.append(f"{tag}: 'env_required' must be an object.")
+    else:
+        for k, v in env_required.items():
+            if not isinstance(k, str):
+                problems.append(f"{tag}: 'env_required' keys must be strings.")
+            if v != "":
+                problems.append(
+                    f"{tag}: env_required[{k!r}] must be an empty string — "
+                    "catalog entries never ship secret values, only the names "
+                    "of env vars the user must fill in."
+                )
+
+    for field in _CATALOG_URL_FIELDS:
+        if field in entry and entry[field] is not None:
+            url = entry[field]
+            if not isinstance(url, str) or not url.startswith("https://"):
+                problems.append(f"{tag}: '{field}' must be an https:// URL.")
+
+    if setup == "link-only":
+        if entry.get("config") is not None:
+            problems.append(f"{tag}: link-only entries must not have a 'config'.")
+        docs_url = entry.get("docs_url")
+        if not isinstance(docs_url, str) or not docs_url.startswith("https://"):
+            problems.append(f"{tag}: link-only entries require an https:// 'docs_url'.")
+    elif setup == "basic":
+        problems.extend(_validate_catalog_config(tag, entry.get("config")))
+
+    return problems
+
+
+def validate_catalog(data) -> list[str]:
+    """
+    Validate a parsed catalog dict. Returns a list of human-readable
+    problems; an EMPTY list means the catalog is valid.
+
+    A non-empty list means REJECT THE WHOLE FILE, not just the offending
+    entry. There is no per-entry salvage here: a catalog that is invalid in
+    one place is untrusted everywhere, because a caller that tried to keep
+    "the other 19 entries that looked fine" would need its own judgment call
+    about which parts of a failed-validation file to trust — exactly the
+    judgment call this function exists to make once, centrally.
+    """
+    if not isinstance(data, dict):
+        return ["Catalog root must be a JSON object."]
+
+    problems: list[str] = []
+
+    schema = data.get("schema")
+    if not isinstance(schema, int) or isinstance(schema, bool) or schema < 1:
+        problems.append("'schema' must be a positive integer.")
+
+    version = data.get("version")
+    if not isinstance(version, int) or isinstance(version, bool) or version < 1:
+        problems.append("'version' must be a positive integer.")
+
+    servers = data.get("servers")
+    if not isinstance(servers, list):
+        problems.append("'servers' must be a list.")
+        return problems  # nothing else to check without a server list
+
+    seen_ids: set[str] = set()
+    for idx, entry in enumerate(servers):
+        problems.extend(_validate_catalog_entry(idx, entry, seen_ids))
+
+    return problems
+
+
+def verify_catalog_signature(raw: bytes, sig: bytes, pubkeys: list[bytes]) -> bool:
+    """
+    Verify an Ed25519 signature over `raw` catalog bytes.
+
+    The signed message is domain-separated: b"bcc-catalog-v1|" + raw, not
+    raw alone (see _CATALOG_SIG_DOMAIN).
+
+    Returns True if ANY key in `pubkeys` verifies — this is what lets keys
+    rotate without bricking installs still trusting an older key.
+
+    Never raises. An invalid signature, a garbage/wrong-length key, a
+    non-bytes argument, an empty signature — all of it just returns False.
+    Signature verification is exactly the wrong place for an exception to
+    accidentally propagate into a code path that fails open.
+    """
+    if not isinstance(raw, bytes) or not isinstance(sig, (bytes, bytearray)):
+        return False
+    if not sig:
+        return False
+    message = _CATALOG_SIG_DOMAIN + raw
+    for pk in pubkeys or []:
+        try:
+            Ed25519PublicKey.from_public_bytes(bytes(pk)).verify(bytes(sig), message)
+            return True
+        except (InvalidSignature, ValueError, TypeError):
+            continue
+    return False
+
+
+def _verify_catalog_candidate(candidate: tuple[bytes, bytes] | None) -> tuple[dict | None, int]:
+    """Verify+load+validate one (raw, sig) candidate. Returns (None, -1) on any failure."""
+    if not candidate:
+        return None, -1
+    raw, sig = candidate
+    if not verify_catalog_signature(raw, sig, CATALOG_PUBKEYS):
+        return None, -1
+    try:
+        data = load_catalog(raw)
+    except (ValueError, TypeError):
+        return None, -1
+    if validate_catalog(data):
+        return None, -1
+    return data, catalog_version(data)
+
+
+def resolve_catalog(
+    bundled: tuple[bytes, bytes] | None,
+    cached: tuple[bytes, bytes] | None,
+    remote: tuple[bytes, bytes] | None,
+    floor: int = 0,
+) -> dict:
+    """
+    Pick the highest-version catalog among bundled/cached/remote. Each of
+    bundled/cached/remote is either None (unavailable) or an (raw_bytes,
+    signature_bytes) pair.
+
+    🔴 SECURITY: every candidate — including `bundled`, the copy frozen into
+    this binary — is verified against CATALOG_PUBKEYS and re-validated from
+    scratch right here. The bundled catalog gets NO implicit trust. This was
+    a hole in the original design: bundling data/catalog.json as a plain
+    asset would let an unsigned/malformed payload that somehow merged to
+    main ship inside the next release and win the version comparison simply
+    by virtue of being local. Signing (and checking the signature at
+    runtime, every time) closes that.
+
+    `floor` is a pure, caller-supplied lower bound (e.g. a persisted
+    "last accepted version" the GUI can load from disk and pass in) — this
+    function does no storage of its own.
+
+    Anti-rollback / anti-freeze, and WHY they apply to every candidate
+    including the first one evaluated: the previous version of this
+    function only ran these checks `if best_version >= 0`, i.e. once a
+    candidate had already been accepted in this pass. That let the FIRST
+    verified candidate through unconditionally — a signed catalog claiming
+    version=999999999 sailed straight past both guards if it happened to be
+    evaluated first, and rollback protection reset on every call anyway
+    (nothing persisted across restarts). Now both guards are anchored to
+    something that doesn't depend on iteration order:
+
+    - The anti-freeze cap is measured against the BUNDLED catalog's version
+      (verified independently, once), not against "whatever was accepted
+      so far in this loop." Bundled ships inside the binary, so it's the
+      one candidate that isn't attacker-supplied at resolve time — the
+      natural trust anchor. If bundled itself doesn't verify, `floor` is
+      the anchor instead.
+    - The anti-rollback floor is max(floor, bundled's version), so a
+      caller that persists `floor` across restarts gets real rollback
+      protection; a caller that doesn't still gets "never below bundled."
+
+    On a version TIE, the bundled candidate wins over cached/remote (it
+    previously lost ties to whichever candidate happened to be evaluated
+    last, silently preferring remote over bundled at equal version).
+
+    Returns the winning catalog dict, or {} if nothing verified and
+    validated.
+    """
+    bundled_data, bundled_version = _verify_catalog_candidate(bundled)
+
+    anchor = bundled_version if bundled_version >= 0 else floor
+    min_accepted = max(floor, bundled_version if bundled_version >= 0 else 0)
+
+    candidates = (
+        ("bundled", bundled_data, bundled_version),
+        ("cached", *_verify_catalog_candidate(cached)),
+        ("remote", *_verify_catalog_candidate(remote)),
+    )
+
+    best: dict = {}
+    best_version = -1
+    best_is_bundled = False
+
+    for source, data, version in candidates:
+        if data is None:
+            continue
+        if version < min_accepted:
+            continue  # anti-rollback / below the persisted floor
+        if version > anchor + _CATALOG_MAX_VERSION_JUMP:
+            continue  # anti-freeze, capped against the bundled trust anchor
+
+        is_bundled = source == "bundled"
+        better = version > best_version or (
+            version == best_version and is_bundled and not best_is_bundled
+        )
+        if better:
+            best = data
+            best_version = version
+            best_is_bundled = is_bundled
+
+    return best
+
+
+def catalog_entry_to_paste_json(entry: dict) -> dict:
+    """
+    Convert a basic-tier catalog entry into the {name: {command, args, env}}
+    shape parse_pasted_json()/_import_server() already understand, so the
+    (future) catalog picker dialog can feed a selection straight into the
+    existing paste-import path instead of growing a parallel one.
+    """
+    config = entry.get("config") or {}
+    name = entry.get("id") or entry.get("display") or "server"
+    data: dict = {
+        "command": config.get("command", ""),
+        "args": list(config.get("args") or []),
+    }
+    env = config.get("env")
+    if env:
+        data["env"] = dict(env)
+    return {str(name): data}
+
+
+def config_has_unfilled_placeholders(cfg: dict) -> bool:
+    """
+    True if any <PLACEHOLDER>-style token remains anywhere in a server
+    config's command/args/env (the shape produced by
+    catalog_entry_to_paste_json). The GUI uses this to refuse Save until
+    every <ALLOWED_DIR>-style token has been filled in with a real value.
+    """
+    values: list[str] = []
+    cmd = cfg.get("command")
+    if isinstance(cmd, str):
+        values.append(cmd)
+    values.extend(a for a in (cfg.get("args") or []) if isinstance(a, str))
+    env = cfg.get("env") or {}
+    if isinstance(env, dict):
+        values.extend(v for v in env.values() if isinstance(v, str))
+    return any(_PLACEHOLDER_RE.search(v) for v in values)

data/catalog.json

@@ -0,0 +1,454 @@
+{
+  "schema": 1,
+  "version": 1,
+  "updated": "2026-07-12",
+  "signed_at": "2026-07-12T21:35:19Z",
+  "servers": [
+    {
+      "id": "filesystem",
+      "display": "Filesystem",
+      "description": "Lets Claude read, write, search, and edit files in directories you explicitly allow.",
+      "category": "files",
+      "homepage": "https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem",
+      "stars": 85995,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "@modelcontextprotocol/server-filesystem@2026.7.10",
+          "<ALLOWED_DIR>"
+        ]
+      },
+      "placeholders": {
+        "<ALLOWED_DIR>": "Absolute path to a directory the server may read/write. Add more directories as additional args."
+      },
+      "env_required": {},
+      "docs_url": "https://github.com/modelcontextprotocol/servers/tree/main/src/filesystem",
+      "notes": "Part of the official modelcontextprotocol/servers reference monorepo (star count is for the whole repo). Clients that support MCP 'roots' can also grant directories dynamically instead of via args.",
+      "last_release": "2026-07-10"
+    },
+    {
+      "id": "fetch",
+      "display": "Fetch",
+      "description": "Fetches a URL and converts the page to clean markdown so Claude can read web content.",
+      "category": "dev",
+      "homepage": "https://github.com/modelcontextprotocol/servers/tree/main/src/fetch",
+      "stars": 85995,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "mcp-server-fetch@2026.7.10"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/modelcontextprotocol/servers/tree/main/src/fetch",
+      "notes": "Can access local/internal IPs, so treat as a mild security risk on untrusted networks. Add '--ignore-robots-txt' or '--user-agent=...' as extra args if needed.",
+      "last_release": "2026-07-10"
+    },
+    {
+      "id": "memory",
+      "display": "Memory",
+      "description": "Gives Claude a persistent knowledge-graph memory that survives across conversations.",
+      "category": "ai",
+      "homepage": "https://github.com/modelcontextprotocol/servers/tree/main/src/memory",
+      "stars": 85995,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "@modelcontextprotocol/server-memory@2026.7.4"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/modelcontextprotocol/servers/tree/main/src/memory",
+      "notes": "Stores entities/relations as a local JSON knowledge graph (path configurable via MEMORY_FILE_PATH env var).",
+      "last_release": "2026-07-04"
+    },
+    {
+      "id": "sequential-thinking",
+      "display": "Sequential Thinking",
+      "description": "Gives Claude a structured, step-by-step reasoning tool for breaking down complex problems.",
+      "category": "ai",
+      "homepage": "https://github.com/modelcontextprotocol/servers/tree/main/src/sequentialthinking",
+      "stars": 85995,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "@modelcontextprotocol/server-sequential-thinking@2026.7.4"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/modelcontextprotocol/servers/tree/main/src/sequentialthinking",
+      "notes": "Set DISABLE_THOUGHT_LOGGING=true to stop it from printing each thought step to the console.",
+      "last_release": "2026-07-04"
+    },
+    {
+      "id": "git",
+      "display": "Git",
+      "description": "Lets Claude read history, diff, branch, and search a local git repository.",
+      "category": "dev",
+      "homepage": "https://github.com/modelcontextprotocol/servers/tree/main/src/git",
+      "stars": 85995,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "mcp-server-git@2026.7.10",
+          "--repository",
+          "<REPO_PATH>"
+        ]
+      },
+      "placeholders": {
+        "<REPO_PATH>": "Absolute path to the local git repository"
+      },
+      "env_required": {},
+      "docs_url": "https://github.com/modelcontextprotocol/servers/tree/main/src/git",
+      "notes": "Operates on a local checkout only; it does not talk to GitHub/GitLab APIs (pair with the GitHub or GitLab server for that).",
+      "last_release": "2026-07-10"
+    },
+    {
+      "id": "github",
+      "display": "GitHub",
+      "description": "Lets Claude read repos/code, and manage issues, pull requests, and Actions on GitHub.",
+      "category": "code-hosting",
+      "homepage": "https://github.com/github/github-mcp-server",
+      "stars": 30202,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "docker",
+        "args": [
+          "run",
+          "-i",
+          "--rm",
+          "-e",
+          "GITHUB_PERSONAL_ACCESS_TOKEN",
+          "ghcr.io/github/github-mcp-server:v1.0.1"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "GITHUB_PERSONAL_ACCESS_TOKEN": ""
+      },
+      "docs_url": "https://github.com/github/github-mcp-server/blob/main/docs/installation-guides/README.md",
+      "notes": "The old '@modelcontextprotocol/server-github' npm package is discontinued (deprecated April 2025). GitHub now ships a Docker-based local server (requires Docker installed/running) plus a hosted remote server at https://api.githubcopilot.com/mcp/ that supports OAuth or PAT auth without Docker."
+    },
+    {
+      "id": "playwright",
+      "display": "Playwright",
+      "description": "Lets Claude drive a real browser (click, type, navigate, screenshot) using Playwright's accessibility-tree snapshots.",
+      "category": "browser",
+      "homepage": "https://github.com/microsoft/playwright-mcp",
+      "stars": 34000,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "@playwright/mcp@0.0.78"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/microsoft/playwright-mcp#readme",
+      "notes": "Maintained by the Playwright team at Microsoft. Add '--isolated' for a throwaway profile, or '--browser firefox|webkit|msedge' to change engine. A persistent browser profile is used by default so logins carry over between sessions.",
+      "last_release": "2026-07-09"
+    },
+    {
+      "id": "chrome-devtools",
+      "display": "Chrome DevTools",
+      "description": "Lets Claude control Chrome and use real DevTools features: performance traces, network inspection, console logs, screenshots.",
+      "category": "browser",
+      "homepage": "https://github.com/ChromeDevTools/chrome-devtools-mcp",
+      "stars": 45000,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "chrome-devtools-mcp@1.5.0"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/ChromeDevTools/chrome-devtools-mcp#readme",
+      "notes": "Maintained by the Google Chrome DevTools team; only officially supports Google Chrome / Chrome for Testing. Exposes the browser's content to the MCP client, so avoid sensitive sites while connected. Add '--slim --headless' for a minimal 3-tool basic-automation mode.",
+      "last_release": "2026-07-03"
+    },
+    {
+      "id": "postgres",
+      "display": "Postgres MCP Pro",
+      "description": "Lets Claude query, inspect schema, and analyze/tune performance of a PostgreSQL database.",
+      "category": "database",
+      "homepage": "https://github.com/crystaldba/postgres-mcp",
+      "stars": 2400,
+      "official": false,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "postgres-mcp@0.3.0",
+          "--access-mode=restricted"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "DATABASE_URI": ""
+      },
+      "docs_url": "https://github.com/crystaldba/postgres-mcp#readme",
+      "notes": "The official reference Postgres server was archived by the MCP team; this community server (Crystal DBA) is the most capable/most-referenced replacement, adding index tuning and EXPLAIN-plan analysis. Use --access-mode=restricted for read-only/production use. Docker image also available (crystaldba/postgres-mcp). Catalog ships --access-mode=restricted (read-only); switch to unrestricted yourself if you want writes.",
+      "last_release": "2025-05-16"
+    },
+    {
+      "id": "n8n",
+      "display": "n8n",
+      "description": "Build, validate, and deploy n8n workflows with full node documentation for the AI.",
+      "category": "infra",
+      "homepage": "https://github.com/czlonkowski/n8n-mcp",
+      "stars": 22257,
+      "official": false,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "n8n-mcp@2.63.2"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "MCP_MODE": "",
+        "N8N_API_URL": "",
+        "N8N_API_KEY": ""
+      },
+      "docs_url": "https://github.com/czlonkowski/n8n-mcp",
+      "notes": "Set MCP_MODE=stdio (required for Claude Desktop, prevents debug logs from breaking the protocol). N8N_API_URL/N8N_API_KEY are optional — without them you still get full node documentation, validation, and template search; with them you get live workflow create/update/execute against your own n8n instance. A hosted free-tier alternative exists at dashboard.n8n-mcp.com.",
+      "last_release": "2026-07-09"
+    },
+    {
+      "id": "notion",
+      "display": "Notion",
+      "description": "Read, search, and edit Notion pages, databases, and comments from your AI assistant.",
+      "category": "productivity",
+      "homepage": "https://github.com/makenotion/notion-mcp-server",
+      "stars": 4400,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "@notionhq/notion-mcp-server@2.4.1"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "NOTION_TOKEN": ""
+      },
+      "docs_url": "https://developers.notion.com/docs/mcp",
+      "notes": "Notion is prioritizing its hosted remote MCP (OAuth, https://mcp.notion.com/mcp) and may eventually sunset this local package, but the stdio server still works today and is the simplest way to get a static config with an internal-integration token.",
+      "last_release": "2026-06-22"
+    },
+    {
+      "id": "obsidian",
+      "display": "Obsidian",
+      "description": "Read, search, and edit notes in your Obsidian vault.",
+      "category": "personal",
+      "homepage": "https://github.com/MarkusPfundstein/mcp-obsidian",
+      "stars": 4067,
+      "official": false,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "mcp-obsidian@0.2.2"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "OBSIDIAN_API_KEY": "",
+        "OBSIDIAN_HOST": "",
+        "OBSIDIAN_PORT": ""
+      },
+      "docs_url": "https://github.com/MarkusPfundstein/mcp-obsidian",
+      "notes": "Requires the Obsidian Local REST API community plugin installed and enabled in Obsidian; copy the API key from the plugin settings. OBSIDIAN_HOST defaults to 127.0.0.1 and OBSIDIAN_PORT to 27124 if omitted.",
+      "last_release": "2025-04-01"
+    },
+    {
+      "id": "brave-search",
+      "display": "Brave Search",
+      "description": "Search the web, news, images, and videos using Brave's independent search index.",
+      "category": "search",
+      "homepage": "https://github.com/brave/brave-search-mcp-server",
+      "stars": 1288,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "@brave/brave-search-mcp-server@2.0.85",
+          "--transport",
+          "stdio"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "BRAVE_API_KEY": ""
+      },
+      "docs_url": "https://github.com/brave/brave-search-mcp-server",
+      "notes": "Official Brave server; replaced the old archived modelcontextprotocol/servers brave-search entry (now in modelcontextprotocol/servers-archived). Get an API key from the Brave Search API dashboard.",
+      "last_release": "2026-06-15"
+    },
+    {
+      "id": "tavily",
+      "display": "Tavily",
+      "description": "AI-optimized web search, extract, map, and crawl API built for LLM agents.",
+      "category": "search",
+      "homepage": "https://github.com/tavily-ai/tavily-mcp",
+      "stars": 2206,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "tavily-mcp@0.2.21"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "TAVILY_API_KEY": ""
+      },
+      "docs_url": "https://github.com/tavily-ai/tavily-mcp",
+      "notes": "Tavily also offers a hosted remote MCP endpoint (mcp.tavily.com) with OAuth as an alternative to running the local npx server.",
+      "last_release": "2026-07-10"
+    },
+    {
+      "id": "home-assistant",
+      "display": "Home Assistant",
+      "description": "Control smart home devices, query states, and troubleshoot automations in Home Assistant.",
+      "category": "smart-home",
+      "homepage": "https://github.com/voska/hass-mcp",
+      "stars": 308,
+      "official": false,
+      "setup": "basic",
+      "config": {
+        "command": "docker",
+        "args": [
+          "run",
+          "-i",
+          "--rm",
+          "-e",
+          "HA_URL",
+          "-e",
+          "HA_TOKEN",
+          "voska/hass-mcp:0.5.0"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {
+        "HA_URL": "",
+        "HA_TOKEN": ""
+      },
+      "docs_url": "https://github.com/voska/hass-mcp",
+      "notes": "HA_URL example: http://homeassistant.local:8123 (use http://host.docker.internal:8123 if HA runs in Docker on the same machine). HA_TOKEN is a Home Assistant long-lived access token from your profile page. A more actively developed alternative is the community 'HA-MCP' integration (homeassistant-ai/ha-mcp, ~3.9k stars), but it installs inside Home Assistant itself via HACS rather than as an external stdio process, so it doesn't fit this catalog's launch-line format."
+    },
+    {
+      "id": "kubernetes",
+      "display": "Kubernetes",
+      "description": "Lets Claude inspect and manage Kubernetes/OpenShift resources — pods, deployments, logs, Helm releases — using your local kubeconfig.",
+      "category": "infra",
+      "homepage": "https://github.com/containers/kubernetes-mcp-server",
+      "stars": 1626,
+      "official": false,
+      "setup": "basic",
+      "config": {
+        "command": "npx",
+        "args": [
+          "-y",
+          "kubernetes-mcp-server@0.0.64"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://github.com/containers/kubernetes-mcp-server#readme",
+      "notes": "Not an official Kubernetes SIG project, but a Go-native (no kubectl dependency) implementation maintained under the 'containers' GitHub org (Podman/Red Hat-adjacent) that's widely regarded as the most capable K8s MCP server, supporting Kubernetes and OpenShift. Uses your existing ~/.kube/config automatically; add --read-only to prevent writes.",
+      "last_release": "2026-07-10"
+    },
+    {
+      "id": "aws-api-mcp-server",
+      "display": "AWS API MCP Server (AWS Labs)",
+      "description": "Lets your AI assistant run AWS CLI commands to inspect and manage AWS resources across virtually every AWS service.",
+      "category": "cloud",
+      "homepage": "https://github.com/awslabs/mcp",
+      "stars": 9431,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "awslabs.aws-api-mcp-server@1.3.46"
+        ]
+      },
+      "placeholders": {},
+      "env_required": {},
+      "docs_url": "https://awslabs.github.io/mcp/servers/aws-api-mcp-server",
+      "notes": "AWS credentials are NOT set in this MCP config — configure them beforehand via `aws configure` (or set AWS_API_MCP_PROFILE_NAME to pick a named profile) so boto3's standard credential chain can find them. Optional env vars: AWS_REGION (default us-east-1), READ_OPERATIONS_ONLY=true to block all mutating AWS calls. AWS notes this server is being superseded by a newer unified AWS MCP server referenced in their agent-toolkit docs.",
+      "last_release": "2026-06-25"
+    },
+    {
+      "id": "grafana",
+      "display": "Grafana",
+      "description": "Query dashboards, datasources, alerts and incidents in Grafana from your AI assistant.",
+      "category": "observability",
+      "homepage": "https://github.com/grafana/mcp-grafana",
+      "stars": 3227,
+      "official": true,
+      "setup": "basic",
+      "config": {
+        "command": "uvx",
+        "args": [
+          "mcp-grafana@0.17.1"
+        ],
+        "env": {
+          "GRAFANA_URL": "<GRAFANA_URL>"
+        }
+      },
+      "placeholders": {
+        "<GRAFANA_URL>": "Your Grafana instance URL, e.g. http://localhost:3000 or https://yourstack.grafana.net"
+      },
+      "env_required": {
+        "GRAFANA_SERVICE_ACCOUNT_TOKEN": ""
+      },
+      "docs_url": "https://grafana.com/docs/grafana/latest/developer-resources/mcp/",
+      "notes": "Requires Grafana 9.0+ for full functionality — datasource-related tools may not work correctly on older versions.",
+      "last_release": "2026-07-07"
+    },
+    {
+      "id": "slack",
+      "display": "Slack",
+      "description": "Search Slack messages, files, and members, and send messages or manage channels from your AI assistant.",
+      "category": "communication",
+      "homepage": "https://slack.com/help/articles/48855576908307-Guide-to-the-Slack-MCP-server",
+      "stars": null,
+      "official": true,
+      "setup": "link-only",
+      "env_required": {},
+      "docs_url": "https://docs.slack.dev/changelog/2026/02/17/slack-mcp/",
+      "notes": "Slack's own MCP server went GA Feb 17, 2026 (streamable HTTP at https://mcp.slack.com/mcp, OAuth). No stdio one-liner is published because it's a hosted, permissioned connector. A well-known community alternative, korotovsky/slack-mcp-server (~1.6k GitHub stars, MIT, not an official Slack product), supports stdio/SSE/HTTP with bot or browser-session tokens and no app-install requirement if a stdio option is preferred."
+    }
+  ]
+}

data/catalog.json.sig

@@ -0,0 +1,2 @@
+ы<8¶ђt2ішл„»‰/НЕ0Тjcw&`
+тrH«MўК›єrBL,0AS€!Х2–иже.SТ°ч–'Agm

pyproject.toml

@@ -1,12 +1,13 @@
 [project]
 name = "better-claude-config"
-version = "1.2.1"
+version = "1.3.0"
 description = "Cross-platform GUI for editing the mcpServers block of Claude Desktop and Claude Code configs"
 readme = "README.md"
 license = { file = "LICENSE" }
 requires-python = ">=3.10"
 dependencies = [
     "PySide6>=6.6",
+    "cryptography>=42.0",
 ]
 
 [project.optional-dependencies]

requirements-dev.txt

@@ -8,3 +8,4 @@ pillow>=10.0        # generates icons/app.ico during CI (Windows build)
 # Test / lint
 pytest>=8.0
 ruff>=0.6
+cryptography>=42.0   # release checksum signing (scripts/sign_checksums.py)

scripts/sign_checksums.py

@@ -0,0 +1,271 @@
+#!/usr/bin/env python3
+"""
+Generate a SHA256SUMS file for release artifacts and sign it with Ed25519.
+
+BCC ships PyInstaller binaries that are not code-signed (no budget for a
+macOS Developer ID / Windows Authenticode certificate). This script provides
+the free half of supply-chain integrity: a checksum manifest, detached-signed
+so downloaders can verify the file they got is the file we published.
+
+This does NOT remove Gatekeeper/SmartScreen warnings and does NOT prove the
+binary is safe to run -- only that it matches what the release signing key
+attested to.
+
+Usage:
+    # Hash every file in a directory into a SHA256SUMS-format manifest.
+    python scripts/sign_checksums.py generate <dir> --out SHA256SUMS
+
+    # Sign a manifest, producing a detached signature.
+    # Private key comes from $RELEASE_SIGNING_KEY (base64 raw Ed25519 seed)
+    # unless --key-b64 is given explicitly (mostly for tests).
+    python scripts/sign_checksums.py sign --sums SHA256SUMS --out SHA256SUMS.sig
+
+    # Verify a manifest against a detached signature and a public key.
+    python scripts/sign_checksums.py verify --sums SHA256SUMS --sig SHA256SUMS.sig \
+        --pubkey-b64 <base64 raw Ed25519 public key>
+
+The private key is generated and rotated via the Catalog Console (#62) --
+this script never generates or stores a key itself.
+"""
+
+from __future__ import annotations
+
+import argparse
+import base64
+import hashlib
+import os
+import sys
+from pathlib import Path
+
+# Domain separation prefix: ties every signature to "a BCC release checksum
+# manifest" so a signature can never be replayed against an unrelated
+# message signed by the same key.
+DOMAIN_PREFIX = b"bcc-release-v1|"
+
+# Public half of the RELEASE signing key(s) -- a SEPARATE keypair from
+# bcc_core.CATALOG_PUBKEYS (issue #68 finding 5). The catalog key is the
+# offline, Console-only root of trust for what BCC executes; this key is
+# CI-resident and signs ONLY the release SHA256SUMS manifest, never the
+# catalog. Keeping them apart means a CI/repo-secret compromise burns the
+# release key -- annoying, but it never lets an attacker sign a catalog a
+# user's binary would trust. A LIST (not a single key), mirroring
+# CATALOG_PUBKEYS, so the release key can be rotated without invalidating
+# the signature on every past release: verification accepts a match against
+# ANY key here.
+#
+# Populated by the maintainer via:
+#     python catalog_console.py keygen --release
+# Rotated 2026-07 (issue #68 finding 5 / #68 CI-exposure incident): the
+# original key was shared with the catalog key and had been exposed to CI,
+# so both keypairs were regenerated as separate, disjoint keys. This list
+# holds only the current release key -- if release.yml's signing-smoke-test
+# ever sees this list empty, it fails closed (loudly) rather than silently
+# verifying against nothing.
+RELEASE_PUBKEYS: list[bytes] = [
+    base64.b64decode("6BnPgJEHJFyVltFoLTCNadIsehjy00iiW8IRlC1TfhA="),
+]
+
+CHUNK_SIZE = 1024 * 1024
+
+
+def sha256_file(path: Path) -> str:
+    """Return the lowercase hex SHA-256 digest of a file's contents."""
+    digest = hashlib.sha256()
+    with open(path, "rb") as fh:
+        while chunk := fh.read(CHUNK_SIZE):
+            digest.update(chunk)
+    return digest.hexdigest()
+
+
+def build_checksums_text(files: dict[str, str]) -> str:
+    """Build a sha256sum(1)-compatible manifest body.
+
+    `files` maps filename -> hex digest. Entries are sorted by filename for
+    a deterministic, diffable output. Format matches `sha256sum` exactly:
+    "<hash>  <filename>\n" (two spaces, no path components).
+    """
+    lines = [f"{digest}  {name}" for name, digest in sorted(files.items())]
+    body = "\n".join(lines)
+    return body + "\n" if body else ""
+
+
+def generate_checksums(directory: Path, *, exclude: set[str] | None = None) -> str:
+    """Hash every regular file directly inside `directory` (non-recursive)
+    and return the SHA256SUMS text. Filenames are recorded without any
+    directory prefix so the manifest can be verified from inside the
+    directory it describes.
+    """
+    exclude = exclude or set()
+    files: dict[str, str] = {}
+    for entry in sorted(directory.iterdir()):
+        if not entry.is_file():
+            continue
+        if entry.name in exclude:
+            continue
+        files[entry.name] = sha256_file(entry)
+    return build_checksums_text(files)
+
+
+def _signing_message(sums_text: str) -> bytes:
+    """The exact bytes that get signed: the domain prefix followed by the
+    raw bytes of the SHA256SUMS file content."""
+    return DOMAIN_PREFIX + sums_text.encode("utf-8")
+
+
+def sign_checksums(seed_b64: str, sums_text: str) -> bytes:
+    """Sign `sums_text` with the Ed25519 private key encoded (base64, raw
+    32-byte seed) in `seed_b64`. Returns the raw 64-byte signature."""
+    # Imported lazily so `generate` mode (used on every CI run) never
+    # requires the `cryptography` package to be installed.
+    from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PrivateKey
+
+    seed = base64.b64decode(seed_b64)
+    if len(seed) != 32:
+        raise ValueError(f"expected a 32-byte raw Ed25519 seed, got {len(seed)} bytes")
+    private_key = Ed25519PrivateKey.from_private_bytes(seed)
+    return private_key.sign(_signing_message(sums_text))
+
+
+def verify_checksums(pubkey_b64: str, sums_text: str, signature: bytes) -> bool:
+    """Verify `signature` over `sums_text` against the base64-encoded raw
+    32-byte Ed25519 public key. Returns True/False; never raises for a bad
+    signature (only for malformed inputs)."""
+    from cryptography.exceptions import InvalidSignature
+    from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PublicKey
+
+    pubkey_bytes = base64.b64decode(pubkey_b64)
+    if len(pubkey_bytes) != 32:
+        raise ValueError(
+            f"expected a 32-byte raw Ed25519 public key, got {len(pubkey_bytes)} bytes"
+        )
+    public_key = Ed25519PublicKey.from_public_bytes(pubkey_bytes)
+    try:
+        public_key.verify(signature, _signing_message(sums_text))
+        return True
+    except InvalidSignature:
+        return False
+
+
+def public_key_b64_from_seed(seed_b64: str) -> str:
+    """Derive the base64 raw public key from a base64 raw seed. Handy for
+    local key-pair sanity checks; not used by the release workflow."""
+    from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PrivateKey
+    from cryptography.hazmat.primitives.serialization import Encoding, PublicFormat
+
+    seed = base64.b64decode(seed_b64)
+    private_key = Ed25519PrivateKey.from_private_bytes(seed)
+    raw = private_key.public_key().public_bytes(Encoding.Raw, PublicFormat.Raw)
+    return base64.b64encode(raw).decode("ascii")
+
+
+def verify_checksums_against_any(pubkeys: list[bytes], sums_text: str, signature: bytes) -> bool:
+    """Verify `signature` against ANY key in `pubkeys` (each a raw 32-byte
+    Ed25519 public key). Mirrors bcc_core.verify_catalog_signature's
+    rotation-friendly "any currently-trusted key" semantics, applied to
+    RELEASE_PUBKEYS instead of the catalog's key list. Returns False (never
+    raises) for an empty `pubkeys` list -- fails closed rather than
+    vacuously verifying against nothing."""
+    return any(
+        verify_checksums(base64.b64encode(pk).decode("ascii"), sums_text, signature)
+        for pk in pubkeys
+    )
+
+
+# --------------------------------------------------------------------------- #
+# CLI
+# --------------------------------------------------------------------------- #
+def _cmd_generate(args: argparse.Namespace) -> int:
+    directory = Path(args.directory)
+    exclude = {"SHA256SUMS", "SHA256SUMS.sig"}
+    text = generate_checksums(directory, exclude=exclude)
+    out_path = Path(args.out)
+    out_path.write_text(text, encoding="utf-8")
+    print(f"Wrote {out_path} ({len(text.splitlines())} entries)")
+    return 0
+
+
+def _cmd_sign(args: argparse.Namespace) -> int:
+    seed_b64 = args.key_b64 or os.environ.get(args.key_env, "")
+    if not seed_b64:
+        print(
+            f"error: no signing key provided (checked --key-b64 and ${args.key_env})",
+            file=sys.stderr,
+        )
+        return 1
+    sums_text = Path(args.sums).read_text(encoding="utf-8")
+    signature = sign_checksums(seed_b64, sums_text)
+    Path(args.out).write_bytes(signature)
+    print(f"Wrote {args.out} ({len(signature)} bytes)")
+    return 0
+
+
+def _cmd_verify(args: argparse.Namespace) -> int:
+    pubkey_b64 = args.pubkey_b64 or os.environ.get(args.pubkey_env, "")
+    if not pubkey_b64:
+        print(
+            f"error: no public key provided (checked --pubkey-b64 and ${args.pubkey_env})",
+            file=sys.stderr,
+        )
+        return 1
+    sums_text = Path(args.sums).read_text(encoding="utf-8")
+    signature = Path(args.sig).read_bytes()
+    ok = verify_checksums(pubkey_b64, sums_text, signature)
+    if ok:
+        print("OK: signature is valid")
+        return 0
+    print("FAILED: signature is invalid", file=sys.stderr)
+    return 1
+
+
+def build_parser() -> argparse.ArgumentParser:
+    parser = argparse.ArgumentParser(
+        description=__doc__, formatter_class=argparse.RawDescriptionHelpFormatter
+    )
+    sub = parser.add_subparsers(dest="mode", required=True)
+
+    p_gen = sub.add_parser(
+        "generate", help="hash every file in a directory into a SHA256SUMS manifest"
+    )
+    p_gen.add_argument("directory", help="directory whose files should be hashed (non-recursive)")
+    p_gen.add_argument("--out", required=True, help="path to write the SHA256SUMS manifest to")
+    p_gen.set_defaults(func=_cmd_generate)
+
+    p_sign = sub.add_parser("sign", help="detached-sign a SHA256SUMS manifest with Ed25519")
+    p_sign.add_argument("--sums", required=True, help="path to the SHA256SUMS manifest to sign")
+    p_sign.add_argument("--out", required=True, help="path to write the detached signature to")
+    p_sign.add_argument(
+        "--key-b64", default=None, help="base64 raw Ed25519 seed (overrides --key-env)"
+    )
+    p_sign.add_argument(
+        "--key-env",
+        default="RELEASE_SIGNING_KEY",
+        help="environment variable holding the base64 seed (default: RELEASE_SIGNING_KEY)",
+    )
+    p_sign.set_defaults(func=_cmd_sign)
+
+    p_verify = sub.add_parser(
+        "verify", help="verify a SHA256SUMS manifest against a detached signature"
+    )
+    p_verify.add_argument("--sums", required=True, help="path to the SHA256SUMS manifest")
+    p_verify.add_argument("--sig", required=True, help="path to the detached signature")
+    p_verify.add_argument(
+        "--pubkey-b64", default=None, help="base64 raw Ed25519 public key (overrides --pubkey-env)"
+    )
+    p_verify.add_argument(
+        "--pubkey-env",
+        default="RELEASE_SIGNING_PUBKEY",
+        help="environment variable holding the base64 public key (default: RELEASE_SIGNING_PUBKEY)",
+    )
+    p_verify.set_defaults(func=_cmd_verify)
+
+    return parser
+
+
+def main(argv: list[str] | None = None) -> int:
+    parser = build_parser()
+    args = parser.parse_args(argv)
+    return args.func(args)
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())

tests/test_catalog_console_git.py

@@ -0,0 +1,215 @@
+"""Tests for catalog_console.py's non-Qt git plumbing and ref-resolution
+seam (issue #68 rotation-completability fix).
+
+catalog_console.py is importable here WITHOUT PySide6 -- its Qt import is
+guarded (`_PYSIDE6_AVAILABLE`) precisely so `keygen`, `show-seed-b64`,
+`keys`, and this git plumbing stay usable (and testable) wherever PySide6
+isn't installed, including this CI test job, which never installs it. If
+PySide6 genuinely isn't importable in this environment, that itself
+exercises the guard path -- see test_module_imports_without_pyside6.
+"""
+
+from __future__ import annotations
+
+import subprocess
+import sys
+from pathlib import Path
+
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent))
+
+import catalog_console as cc
+import catalog_review as review
+
+_SEED_CATALOG = b'{"schema": 1, "version": 1, "servers": []}'
+_SEED_SIG = b"\x00" * 64
+
+
+def _run(*args: str, cwd: Path) -> None:
+    subprocess.run(["git", *args], cwd=cwd, check=True, capture_output=True)
+
+
+def _init_bare_and_clone(tmp_path: Path) -> tuple[Path, Path]:
+    """A bare "origin" repo with `main` and `rotation-branch` both seeded
+    with a catalog + (dummy) signature, plus a working clone with `origin`
+    already configured -- mirroring the tokened-remote clone
+    catalog_console.py's git plumbing is always run against."""
+    origin = tmp_path / "origin.git"
+    _run("init", "--bare", str(origin), cwd=tmp_path)
+
+    seed = tmp_path / "seed"
+    _run("clone", str(origin), str(seed), cwd=tmp_path)
+    _run("config", "user.email", "test@example.com", cwd=seed)
+    _run("config", "user.name", "Test", cwd=seed)
+
+    (seed / "data").mkdir()
+    (seed / "data" / "catalog.json").write_bytes(_SEED_CATALOG)
+    (seed / "data" / "catalog.json.sig").write_bytes(_SEED_SIG)
+    _run("add", "-A", cwd=seed)
+    _run("commit", "-m", "seed", cwd=seed)
+    _run("push", "origin", "HEAD:refs/heads/main", cwd=seed)
+    _run("checkout", "-b", "rotation-branch", cwd=seed)
+    _run("push", "origin", "HEAD:refs/heads/rotation-branch", cwd=seed)
+
+    clone = tmp_path / "work"
+    _run("clone", str(origin), str(clone), cwd=tmp_path)
+    _run("config", "user.email", "test@example.com", cwd=clone)
+    _run("config", "user.name", "Test", cwd=clone)
+    return origin, clone
+
+
+# --------------------------------------------------------------------------- #
+# The module must stay importable without PySide6 -- this IS the fix that
+# lets `keys`/`keygen`/`show-seed-b64` (and this whole test file) run
+# somewhere PySide6 isn't installed.
+# --------------------------------------------------------------------------- #
+def test_module_imports_without_pyside6():
+    assert hasattr(cc, "_PYSIDE6_AVAILABLE")
+    # This CI test job never installs PySide6 (see .github/workflows/ci.yml
+    # "Install test dependencies": pytest + cryptography only) -- so on CI,
+    # this assertion is itself proof the guard is doing its job. Locally,
+    # where a maintainer's env DOES have PySide6, it's fine either way; the
+    # only real assertion this test needs is "importing the module never
+    # raises", which happened just by getting this far.
+    assert cc._PYSIDE6_AVAILABLE in (True, False)
+
+
+def test_cmd_gui_fails_soft_without_pyside6(monkeypatch, capsys):
+    if cc._PYSIDE6_AVAILABLE:
+        return  # nothing to prove where PySide6 IS available
+    import argparse
+
+    args = argparse.Namespace(repo=".", ref=None)
+    assert cc.cmd_gui(args) == 1
+    assert "PySide6" in capsys.readouterr().err
+
+
+# --------------------------------------------------------------------------- #
+# compute_own_refs: the PURE ref-resolution seam. No git, no Qt.
+# --------------------------------------------------------------------------- #
+def test_compute_own_refs_defaults_to_main_only():
+    assert cc.compute_own_refs(None, None) == ["main"]
+
+
+def test_compute_own_refs_adds_detected_branch():
+    assert cc.compute_own_refs(None, "chore/68-key-rotation") == [
+        "main",
+        "chore/68-key-rotation",
+    ]
+
+
+def test_compute_own_refs_explicit_ref_overrides_detected_branch():
+    assert cc.compute_own_refs("explicit-branch", "detected-branch") == [
+        "main",
+        "explicit-branch",
+    ]
+
+
+def test_compute_own_refs_does_not_duplicate_main():
+    assert cc.compute_own_refs(None, "main") == ["main"]
+    assert cc.compute_own_refs("main", "some-other-branch") == ["main"]
+
+
+# --------------------------------------------------------------------------- #
+# current_branch: git plumbing, no Qt.
+# --------------------------------------------------------------------------- #
+def test_current_branch_detects_checked_out_branch(tmp_path):
+    _origin, clone = _init_bare_and_clone(tmp_path)
+    _run("fetch", "origin", "rotation-branch", cwd=clone)
+    _run("checkout", "-B", "rotation-branch", "origin/rotation-branch", cwd=clone)
+    assert cc.current_branch(clone) == "rotation-branch"
+
+
+def test_current_branch_none_on_detached_head(tmp_path):
+    _origin, clone = _init_bare_and_clone(tmp_path)
+    commit = cc.fetch_ref(clone, "main")
+    _run("checkout", commit, cwd=clone)
+    assert cc.current_branch(clone) is None
+
+
+# --------------------------------------------------------------------------- #
+# commit_and_push_signed_catalog: MUST target the given branch, never a
+# hardcoded "main" -- issue #68's completability fix. This is exactly the
+# bug that, before the fix, would have made ReviewWindow._on_sign push a
+# PR/branch review's signature straight to main regardless of what was
+# actually reviewed.
+# --------------------------------------------------------------------------- #
+def test_commit_and_push_signed_catalog_targets_the_given_branch_not_main(tmp_path):
+    _origin, clone = _init_bare_and_clone(tmp_path)
+
+    new_raw = b'{"schema": 1, "version": 2, "servers": []}'
+    new_sig = b"\x01" * 64
+    cc.commit_and_push_signed_catalog(clone, new_raw, new_sig, branch="rotation-branch")
+
+    rotation_commit = cc.fetch_ref(clone, "rotation-branch")
+    rotation_raw, _sha = cc.read_catalog_at_commit(clone, rotation_commit)
+    assert rotation_raw == new_raw
+
+    # main on the shared origin must be COMPLETELY untouched by a sign that
+    # was reviewed and pushed against rotation-branch.
+    main_commit = cc.fetch_ref(clone, "main")
+    main_raw, _sha = cc.read_catalog_at_commit(clone, main_commit)
+    assert main_raw == _SEED_CATALOG
+
+
+def test_commit_and_push_signed_catalog_still_defaults_to_main(tmp_path):
+    """Backward-compatible default: callers that don't pass `branch` (there
+    are none left in catalog_console.py itself, but the signature keeps the
+    default for any other caller / test fixture) still push to main."""
+    _origin, clone = _init_bare_and_clone(tmp_path)
+
+    new_raw = b'{"schema": 1, "version": 2, "servers": []}'
+    new_sig = b"\x01" * 64
+    cc.commit_and_push_signed_catalog(clone, new_raw, new_sig)
+
+    main_commit = cc.fetch_ref(clone, "main")
+    main_raw, _sha = cc.read_catalog_at_commit(clone, main_commit)
+    assert main_raw == new_raw
+
+    rotation_commit = cc.fetch_ref(clone, "rotation-branch")
+    rotation_raw, _sha = cc.read_catalog_at_commit(clone, rotation_commit)
+    assert rotation_raw == _SEED_CATALOG  # untouched
+
+
+# --------------------------------------------------------------------------- #
+# catalog_sig_status_on_disk: the check behind `keys`' "does catalog.json.sig
+# currently verify?" line -- this is precisely the check that would have
+# caught the current chore/68-key-rotation state (bcc_core.CATALOG_PUBKEYS
+# rotated, data/catalog.json.sig still signed by the retired key).
+# --------------------------------------------------------------------------- #
+def test_catalog_sig_status_on_disk_valid(tmp_path):
+    seed, pub = review.generate_keypair()
+    raw = b'{"schema": 1, "version": 1, "servers": []}'
+    sig = review.sign_catalog_bytes(raw, seed)
+
+    (tmp_path / "data").mkdir()
+    (tmp_path / "data" / "catalog.json").write_bytes(raw)
+    (tmp_path / "data" / "catalog.json.sig").write_bytes(sig)
+
+    assert cc.catalog_sig_status_on_disk(tmp_path, [pub]) == "valid"
+
+
+def test_catalog_sig_status_on_disk_invalid_when_pubkey_rotated(tmp_path):
+    """The exact chore/68-key-rotation scenario: signed by an OLD key, but
+    the committed pubkey list now only has the NEW key."""
+    old_seed, _old_pub = review.generate_keypair()
+    _new_seed, new_pub = review.generate_keypair()
+    raw = b'{"schema": 1, "version": 1, "servers": []}'
+    sig = review.sign_catalog_bytes(raw, old_seed)
+
+    (tmp_path / "data").mkdir()
+    (tmp_path / "data" / "catalog.json").write_bytes(raw)
+    (tmp_path / "data" / "catalog.json.sig").write_bytes(sig)
+
+    assert cc.catalog_sig_status_on_disk(tmp_path, [new_pub]) == "invalid"
+
+
+def test_catalog_sig_status_on_disk_missing_when_no_sig_file(tmp_path):
+    (tmp_path / "data").mkdir()
+    (tmp_path / "data" / "catalog.json").write_bytes(b"{}")
+    assert cc.catalog_sig_status_on_disk(tmp_path, []) == "missing"
+
+
+def test_catalog_sig_status_on_disk_missing_when_no_catalog_file(tmp_path):
+    (tmp_path / "data").mkdir()
+    (tmp_path / "data" / "catalog.json.sig").write_bytes(b"\x00" * 64)
+    assert cc.catalog_sig_status_on_disk(tmp_path, []) == "missing"

tests/test_catalog_console_packaging.py

@@ -0,0 +1,66 @@
+"""Asserts the maintainer-only Catalog Console (catalog_console.py,
+catalog_review.py) is never bundled into the release binary.
+
+A signing/review tool shipping to end users would be an own-goal (issue
+#62): it has no reason to run on a user's machine, and its presence would
+be a confusing artefact of a build that's supposed to be a thin GUI over
+mcpServers config editing."""
+
+from __future__ import annotations
+
+from pathlib import Path
+
+REPO_ROOT = Path(__file__).resolve().parent.parent
+SPEC_PATH = REPO_ROOT / "bcc.spec"
+
+_EXCLUDED_FILES = ("catalog_console.py", "catalog_review.py")
+
+
+def test_spec_file_exists():
+    assert SPEC_PATH.exists()
+
+
+def test_console_files_not_named_in_spec():
+    """The spec text must never reference either maintainer-only module --
+    not as the Analysis entry point, not in datas, not anywhere."""
+    spec_text = SPEC_PATH.read_text(encoding="utf-8")
+    for filename in _EXCLUDED_FILES:
+        assert filename not in spec_text, (
+            f"{filename} must never be referenced by bcc.spec -- it is a "
+            "maintainer-only tool and must not ship to users."
+        )
+
+
+def test_spec_analysis_entry_point_is_bcc_py_only():
+    """PyInstaller's Analysis(...) call determines the dependency-scanned
+    entry point(s); it must be bcc.py alone."""
+    spec_text = SPEC_PATH.read_text(encoding="utf-8")
+    assert 'Analysis(\n    ["bcc.py"],' in spec_text or 'Analysis(["bcc.py"]' in spec_text, (
+        "bcc.spec's Analysis(...) entry point changed shape -- re-verify by hand "
+        "that catalog_console.py / catalog_review.py are still excluded."
+    )
+
+
+def test_console_modules_exist_but_are_standalone_top_level_files():
+    """Sanity check the files this test is guarding actually exist as
+    top-level modules (not, say, silently moved into a package PyInstaller's
+    Analysis would still pick up as an implicit import of bcc.py)."""
+    for filename in _EXCLUDED_FILES:
+        assert (REPO_ROOT / filename).exists()
+        # bcc.py must not import them.
+        bcc_text = (REPO_ROOT / "bcc.py").read_text(encoding="utf-8")
+        module_name = filename.removesuffix(".py")
+        assert f"import {module_name}" not in bcc_text
+        assert f"from {module_name}" not in bcc_text
+
+
+def test_requirements_files_do_not_reference_console_only_needs():
+    """catalog_console.py's only import beyond the shipped stack is the
+    optional `keyring` package, which is intentionally NOT added as a hard
+    dependency anywhere a user install would pick it up."""
+    for req_file in ("requirements.txt", "requirements-dev.txt"):
+        path = REPO_ROOT / req_file
+        if not path.exists():
+            continue
+        text = path.read_text(encoding="utf-8").lower()
+        assert "keyring" not in text

tests/test_checksums.py

@@ -0,0 +1,234 @@
+"""Tests for scripts/sign_checksums.py: SHA256SUMS generation and detached
+Ed25519 signing/verification for release artifacts."""
+
+from __future__ import annotations
+
+import base64
+import subprocess
+import sys
+from pathlib import Path
+
+import pytest
+
+sys.path.insert(0, str(Path(__file__).resolve().parent.parent / "scripts"))
+
+import sign_checksums as sc
+
+cryptography = pytest.importorskip("cryptography")
+from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PrivateKey  # noqa: E402
+from cryptography.hazmat.primitives.serialization import Encoding, PublicFormat  # noqa: E402
+
+
+def _make_keypair() -> tuple[str, str]:
+    """Return (seed_b64, pubkey_b64) for a fresh Ed25519 keypair."""
+    private_key = Ed25519PrivateKey.generate()
+    seed = private_key.private_bytes_raw()
+    pubkey = private_key.public_key().public_bytes(Encoding.Raw, PublicFormat.Raw)
+    return base64.b64encode(seed).decode("ascii"), base64.b64encode(pubkey).decode("ascii")
+
+
+# --------------------------------------------------------------------------- #
+# sha256_file / build_checksums_text / generate_checksums
+# --------------------------------------------------------------------------- #
+def test_sha256_file_matches_hashlib(tmp_path):
+    f = tmp_path / "a.txt"
+    f.write_bytes(b"hello world")
+    import hashlib
+
+    assert sc.sha256_file(f) == hashlib.sha256(b"hello world").hexdigest()
+
+
+def test_build_checksums_text_sorted_and_formatted():
+    files = {"zeta.zip": "aa" * 32, "alpha.zip": "bb" * 32}
+    text = sc.build_checksums_text(files)
+    lines = text.splitlines()
+    assert lines[0].endswith("alpha.zip")
+    assert lines[1].endswith("zeta.zip")
+    # Standard sha256sum format: hash, two spaces, filename.
+    assert lines[0] == f"{'bb' * 32}  alpha.zip"
+
+
+def test_build_checksums_text_empty():
+    assert sc.build_checksums_text({}) == ""
+
+
+def test_generate_checksums_from_directory(tmp_path):
+    (tmp_path / "b.bin").write_bytes(b"second")
+    (tmp_path / "a.bin").write_bytes(b"first")
+    (tmp_path / "subdir").mkdir()
+    (tmp_path / "subdir" / "ignored.bin").write_bytes(b"nested, not hashed")
+
+    text = sc.generate_checksums(tmp_path)
+    lines = text.splitlines()
+    assert len(lines) == 2
+    assert lines[0].endswith("a.bin")
+    assert lines[1].endswith("b.bin")
+    assert "subdir" not in text
+
+
+def test_generate_checksums_excludes_manifest_files(tmp_path):
+    (tmp_path / "archive.zip").write_bytes(b"payload")
+    (tmp_path / "SHA256SUMS").write_text("stale")
+    (tmp_path / "SHA256SUMS.sig").write_bytes(b"stale-sig")
+
+    text = sc.generate_checksums(tmp_path, exclude={"SHA256SUMS", "SHA256SUMS.sig"})
+    assert "archive.zip" in text
+    assert "SHA256SUMS" not in text.replace("archive.zip", "")
+
+
+# --------------------------------------------------------------------------- #
+# sign_checksums / verify_checksums
+# --------------------------------------------------------------------------- #
+def test_sign_then_verify_roundtrip():
+    seed_b64, pubkey_b64 = _make_keypair()
+    sums_text = "deadbeef" * 8 + "  BetterClaudeConfig-Linux.tar.gz\n"
+
+    signature = sc.sign_checksums(seed_b64, sums_text)
+    assert len(signature) == 64
+    assert sc.verify_checksums(pubkey_b64, sums_text, signature) is True
+
+
+def test_verify_rejects_tampered_checksums():
+    seed_b64, pubkey_b64 = _make_keypair()
+    sums_text = "aa" * 32 + "  file.zip\n"
+    signature = sc.sign_checksums(seed_b64, sums_text)
+
+    tampered = "bb" * 32 + "  file.zip\n"
+    assert sc.verify_checksums(pubkey_b64, tampered, signature) is False
+
+
+def test_verify_rejects_wrong_key():
+    seed_b64, _ = _make_keypair()
+    _, other_pubkey_b64 = _make_keypair()
+    sums_text = "cc" * 32 + "  file.zip\n"
+    signature = sc.sign_checksums(seed_b64, sums_text)
+
+    assert sc.verify_checksums(other_pubkey_b64, sums_text, signature) is False
+
+
+def test_domain_prefix_is_applied():
+    """The signed message must be prefixed, not the raw manifest bytes --
+    otherwise a signature over this manifest could be replayed as a
+    signature over an unrelated message with the same bytes elsewhere."""
+    seed_b64, pubkey_b64 = _make_keypair()
+    sums_text = "11" * 32 + "  file.zip\n"
+
+    from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PrivateKey as PK
+
+    seed = base64.b64decode(seed_b64)
+    raw_signature = PK.from_private_bytes(seed).sign(sums_text.encode("utf-8"))
+
+    # A signature over the raw (unprefixed) bytes must NOT verify via our
+    # domain-separated verify function.
+    assert sc.verify_checksums(pubkey_b64, sums_text, raw_signature) is False
+
+    # But our own sign_checksums() output does verify.
+    good_signature = sc.sign_checksums(seed_b64, sums_text)
+    assert sc.verify_checksums(pubkey_b64, sums_text, good_signature) is True
+
+
+def test_sign_checksums_rejects_bad_seed_length():
+    bad_seed_b64 = base64.b64encode(b"too-short").decode("ascii")
+    with pytest.raises(ValueError):
+        sc.sign_checksums(bad_seed_b64, "irrelevant\n")
+
+
+def test_verify_checksums_rejects_bad_pubkey_length():
+    seed_b64, _ = _make_keypair()
+    sig = sc.sign_checksums(seed_b64, "irrelevant\n")
+    bad_pubkey_b64 = base64.b64encode(b"too-short").decode("ascii")
+    with pytest.raises(ValueError):
+        sc.verify_checksums(bad_pubkey_b64, "irrelevant\n", sig)
+
+
+def test_public_key_b64_from_seed_matches_generated_pubkey():
+    seed_b64, pubkey_b64 = _make_keypair()
+    assert sc.public_key_b64_from_seed(seed_b64) == pubkey_b64
+
+
+# --------------------------------------------------------------------------- #
+# CLI (end-to-end, via subprocess so argparse wiring is exercised too)
+# --------------------------------------------------------------------------- #
+SCRIPT = Path(__file__).resolve().parent.parent / "scripts" / "sign_checksums.py"
+
+
+def _run(*args, env=None):
+    return subprocess.run(
+        [sys.executable, str(SCRIPT), *args],
+        capture_output=True,
+        text=True,
+        env=env,
+    )
+
+
+def test_cli_generate_sign_verify_roundtrip(tmp_path, monkeypatch):
+    seed_b64, pubkey_b64 = _make_keypair()
+
+    release_dir = tmp_path / "release-files"
+    release_dir.mkdir()
+    (release_dir / "BetterClaudeConfig-Linux.tar.gz").write_bytes(b"fake archive contents")
+    (release_dir / "BetterClaudeConfig-macOS.zip").write_bytes(b"fake zip contents")
+
+    sums_path = release_dir / "SHA256SUMS"
+    sig_path = release_dir / "SHA256SUMS.sig"
+
+    gen = _run("generate", str(release_dir), "--out", str(sums_path))
+    assert gen.returncode == 0, gen.stderr
+    assert sums_path.exists()
+    body = sums_path.read_text()
+    assert "BetterClaudeConfig-Linux.tar.gz" in body
+    assert "BetterClaudeConfig-macOS.zip" in body
+
+    sign = _run("sign", "--sums", str(sums_path), "--out", str(sig_path), "--key-b64", seed_b64)
+    assert sign.returncode == 0, sign.stderr
+    assert sig_path.exists()
+    assert sig_path.stat().st_size == 64
+
+    verify = _run(
+        "verify",
+        "--sums",
+        str(sums_path),
+        "--sig",
+        str(sig_path),
+        "--pubkey-b64",
+        pubkey_b64,
+    )
+    assert verify.returncode == 0, verify.stderr
+    assert "OK" in verify.stdout
+
+
+def test_cli_sign_without_key_fails_loudly(tmp_path):
+    sums_path = tmp_path / "SHA256SUMS"
+    sums_path.write_text("aa" * 32 + "  file.zip\n")
+    sig_path = tmp_path / "SHA256SUMS.sig"
+
+    import os
+
+    env = {k: v for k, v in os.environ.items() if k != "RELEASE_SIGNING_KEY"}
+    result = _run("sign", "--sums", str(sums_path), "--out", str(sig_path), env=env)
+
+    assert result.returncode != 0
+    assert not sig_path.exists(), "must never write a bogus/empty signature file"
+    assert "no signing key" in result.stderr.lower()
+
+
+def test_cli_verify_detects_tampering(tmp_path):
+    seed_b64, pubkey_b64 = _make_keypair()
+    sums_path = tmp_path / "SHA256SUMS"
+    sums_path.write_text("aa" * 32 + "  file.zip\n")
+    sig_path = tmp_path / "SHA256SUMS.sig"
+
+    _run("sign", "--sums", str(sums_path), "--out", str(sig_path), "--key-b64", seed_b64)
+
+    sums_path.write_text("bb" * 32 + "  file.zip\n")  # tamper after signing
+    verify = _run(
+        "verify",
+        "--sums",
+        str(sums_path),
+        "--sig",
+        str(sig_path),
+        "--pubkey-b64",
+        pubkey_b64,
+    )
+    assert verify.returncode != 0
+    assert "FAILED" in verify.stdout + verify.stderr

tests/test_core.py

@@ -10,6 +10,7 @@ import urllib.request
 from pathlib import Path
 
 import pytest
+from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PrivateKey
 
 import bcc_core as c
 
@@ -67,6 +68,64 @@ def test_legacy_settings_json_surfaced_only_with_servers(fake_home):
     assert any("legacy" in p.label for p in c.discover_profiles())
 
 
+def test_project_with_mcp_json_is_discovered(tmp_path):
+    proj = tmp_path / "my-project"
+    proj.mkdir()
+    (proj / ".mcp.json").write_text('{"mcpServers": {"x": {"command": "npx"}}}')
+    claude_json = tmp_path / ".claude.json"
+    claude_json.write_text(json.dumps({"projects": {str(proj): {}}}))
+
+    profs = c.discover_project_configs(claude_json)
+    assert len(profs) == 1
+    assert profs[0].label == f"Project: {proj.name}"
+    assert profs[0].path == proj / ".mcp.json"
+    assert profs[0].config_exists
+
+
+def test_project_without_mcp_json_not_listed(tmp_path):
+    proj = tmp_path / "no-mcp-project"
+    proj.mkdir()
+    claude_json = tmp_path / ".claude.json"
+    claude_json.write_text(json.dumps({"projects": {str(proj): {}}}))
+
+    assert c.discover_project_configs(claude_json) == []
+
+
+def test_projects_missing_or_not_dict_yields_no_profiles(tmp_path):
+    claude_json = tmp_path / ".claude.json"
+
+    claude_json.write_text(json.dumps({}))
+    assert c.discover_project_configs(claude_json) == []
+
+    claude_json.write_text(json.dumps({"projects": ["not", "a", "dict"]}))
+    assert c.discover_project_configs(claude_json) == []
+
+
+def test_malformed_claude_json_fails_quiet(fake_home):
+    (fake_home / ".claude.json").write_text("{not valid json")
+    assert c.discover_project_configs(fake_home / ".claude.json") == []
+    # discover_profiles as a whole must still work and just skip project profiles
+    profs = c.discover_profiles()
+    assert not any(p.label.startswith("Project:") for p in profs)
+
+
+def test_project_configs_deduped_against_existing_profiles(fake_home):
+    # Point a project directly at the Claude Code profile's own .mcp.json-shaped
+    # path to prove discover_profiles() won't duplicate an already-listed path.
+    proj = fake_home / "dup-project"
+    proj.mkdir()
+    mcp_path = proj / ".mcp.json"
+    mcp_path.write_text('{"mcpServers": {"x": {"command": "npx"}}}')
+    (fake_home / ".claude.json").write_text(json.dumps({"projects": {str(proj): {}}}))
+
+    direct = c.discover_project_configs(fake_home / ".claude.json")
+    assert len(direct) == 1
+
+    profs = c.discover_profiles()
+    project_profiles = [p for p in profs if str(p.path) == str(mcp_path)]
+    assert len(project_profiles) == 1
+
+
 # --------------------------------------------------------------------------- #
 # 2. Write pipeline: preserves other keys + order, only touches mcpServers
 # --------------------------------------------------------------------------- #
@@ -181,6 +240,85 @@ def test_valid_set_passes_clean():
     assert c.validate_servers([c.ServerEntry("good", {"command": "node"}, True)]) == []
 
 
+# --------------------------------------------------------------------------- #
+# 4b. Lint: non-blocking structural warnings
+# --------------------------------------------------------------------------- #
+def test_lint_flags_non_string_command():
+    warnings = c.lint_server("x", {"command": 5})
+    assert any("'command' should be a string" in w for w in warnings)
+
+
+def test_lint_flags_args_not_a_list():
+    warnings = c.lint_server("x", {"command": "node", "args": "-y foo"})
+    assert any("'args' should be a list" in w for w in warnings)
+
+
+def test_lint_flags_args_with_non_string_items():
+    warnings = c.lint_server("x", {"command": "node", "args": ["-y", 5]})
+    assert any("'args' contains non-string values" in w for w in warnings)
+
+
+def test_lint_flags_env_not_a_dict():
+    warnings = c.lint_server("x", {"command": "node", "env": ["FOO=bar"]})
+    assert any("'env' should be an object" in w for w in warnings)
+
+
+def test_lint_flags_env_with_non_string_values():
+    warnings = c.lint_server("x", {"command": "node", "env": {"FOO": 5}})
+    assert any("'env' contains non-string values" in w for w in warnings)
+
+
+def test_lint_flags_headers_not_a_dict():
+    warnings = c.lint_server("x", {"url": "https://x", "headers": ["Authorization: x"]})
+    assert any("'headers' should be an object" in w for w in warnings)
+
+
+def test_lint_flags_headers_with_non_string_values():
+    warnings = c.lint_server("x", {"url": "https://x", "headers": {"Authorization": 5}})
+    assert any("'headers' contains non-string values" in w for w in warnings)
+
+
+def test_lint_flags_bad_type_value():
+    warnings = c.lint_server("x", {"url": "https://x", "type": "websocket"})
+    assert any("'type'" in w and "websocket" in w for w in warnings)
+
+
+def test_lint_flags_extra_unknown_fields():
+    warnings = c.lint_server("x", {"command": "node", "cwd": "/tmp", "timeout": 30})
+    matches = [w for w in warnings if "extra fields preserved as-is" in w]
+    assert len(matches) == 1
+    assert "cwd" in matches[0]
+    assert "timeout" in matches[0]
+
+
+def test_lint_clean_server_yields_no_warnings():
+    assert c.lint_server("good", {"command": "node", "args": ["a.js"]}) == []
+    assert (
+        c.lint_server(
+            "remote", {"url": "https://x", "type": "http", "headers": {"Authorization": "x"}}
+        )
+        == []
+    )
+
+
+def test_lint_servers_aggregates_across_entries():
+    entries = [
+        c.ServerEntry("a", {"command": 5}, True),
+        c.ServerEntry("b", {"url": "https://x", "type": "bogus"}, True),
+    ]
+    warnings = c.lint_servers(entries)
+    assert any("'a'" in w and "'command' should be a string" in w for w in warnings)
+    assert any("'b'" in w and "'type'" in w for w in warnings)
+
+
+def test_lint_warning_is_not_a_blocking_problem():
+    # args given as a single string isn't caught by validate_servers (a
+    # command is still present), but it's a lint warning.
+    entry = c.ServerEntry("x", {"command": "node", "args": "-y foo"}, True)
+    assert c.validate_servers([entry]) == []
+    assert c.lint_servers([entry]) != []
+
+
 # --------------------------------------------------------------------------- #
 # 5. Secrets: detection + redaction
 # --------------------------------------------------------------------------- #
@@ -769,6 +907,68 @@ def test_args_secret_warning_empty():
     assert c.args_secret_warning({"args": []}) is None
 
 
+# --------------------------------------------------------------------------- #
+# Named server sets (issue #52)
+# --------------------------------------------------------------------------- #
+def _three_servers():
+    return [
+        c.ServerEntry("alpha", {"command": "npx"}, True),
+        c.ServerEntry("beta", {"command": "uvx"}, True),
+        c.ServerEntry("gamma", {"url": "https://x.example/mcp"}, False),
+    ]
+
+
+def test_save_and_list_server_sets_roundtrip():
+    cfg: dict = {}
+    members = c.save_server_set(cfg, "webdev", _three_servers())
+    assert members == ["alpha", "beta"]  # only the enabled ones
+    assert c.list_server_sets(cfg) == {"webdev": ["alpha", "beta"]}
+
+
+def test_apply_server_set_enables_exactly_the_members():
+    servers = _three_servers()
+    missing = c.apply_server_set(servers, ["gamma"])
+    assert missing == []
+    assert [s.enabled for s in servers] == [False, False, True]
+
+
+def test_apply_server_set_reports_missing_members():
+    servers = _three_servers()
+    missing = c.apply_server_set(servers, ["alpha", "vanished", "gone"])
+    assert missing == ["gone", "vanished"]
+    assert [s.enabled for s in servers] == [True, False, False]  # rest still applied
+
+
+def test_delete_server_set_drops_empty_key():
+    cfg: dict = {}
+    c.save_server_set(cfg, "only", _three_servers())
+    assert c.delete_server_set(cfg, "only") is True
+    assert c.SETS_KEY not in cfg  # no empty bcc key left behind
+    assert c.delete_server_set(cfg, "only") is False
+
+
+def test_server_sets_survive_write_and_reload(tmp_path):
+    cfgpath = tmp_path / "cfg.json"
+    cfg = {"mcpServers": {"alpha": {"command": "npx"}}}
+    c.save_server_set(cfg, "webdev", [c.ServerEntry("alpha", {"command": "npx"}, True)])
+    c.write_config(cfgpath, cfg)
+    reloaded = c.load_config(cfgpath)
+    assert c.list_server_sets(reloaded) == {"webdev": ["alpha"]}
+
+
+def test_apply_servers_preserves_sets_key():
+    cfg: dict = {"mcpServers": {}}
+    c.save_server_set(cfg, "s", [c.ServerEntry("alpha", {"command": "npx"}, True)])
+    c.apply_servers(cfg, _three_servers())
+    assert c.SETS_KEY in cfg  # the server writer never touches sets
+
+
+def test_list_server_sets_skips_malformed_entries():
+    cfg = {c.SETS_KEY: {"good": ["a"], "bad-not-list": "a", "bad-items": ["a", 3]}}
+    assert c.list_server_sets(cfg) == {"good": ["a"]}
+    assert c.list_server_sets({c.SETS_KEY: "junk"}) == {}
+
+
 # --------------------------------------------------------------------------- #
 # resolve_name_collision (paste/import duplicate-name handling — issue #8)
 # --------------------------------------------------------------------------- #
@@ -1469,3 +1669,704 @@ def test_app_icon_assets_present():
         assert (rounded / f"icon-{size}.png").is_file(), f"missing icon-{size}.png"
     assert (root / "icons" / "app.ico").is_file()
     assert (root / "icons" / "app.icns").is_file()
+
+
+# --------------------------------------------------------------------------- #
+# MCP server catalog (issue #10 / #61)
+# --------------------------------------------------------------------------- #
+def _minimal_catalog(version: int = 1) -> dict:
+    return {
+        "schema": 1,
+        "version": version,
+        "updated": "2026-07-12",
+        "servers": [
+            {
+                "id": "widget",
+                "display": "Widget",
+                "description": "A test widget server.",
+                "category": "dev",
+                "homepage": "https://example.com/widget",
+                "stars": 10,
+                "official": True,
+                "setup": "basic",
+                "config": {
+                    # Pinned on purpose (issue #68 finding 3): an earlier
+                    # version of this fixture used an unpinned package and
+                    # asserted it validated clean, which enshrined the bug
+                    # instead of catching it.
+                    "command": "npx",
+                    "args": ["-y", "widget-mcp@1.0.0"],
+                },
+                "placeholders": {},
+                "env_required": {},
+                "docs_url": "https://example.com/widget/docs",
+                "notes": "",
+            }
+        ],
+    }
+
+
+def _catalog_with(server_overrides: dict) -> dict:
+    data = _minimal_catalog()
+    data["servers"][0].update(server_overrides)
+    return data
+
+
+def _sign(raw: bytes, priv: Ed25519PrivateKey) -> bytes:
+    # Independent of bcc_core's domain-separation constant on purpose: this
+    # is the literal wire format the design calls for, hardcoded here so a
+    # change to the constant would be caught as a real behaviour change.
+    return priv.sign(b"bcc-catalog-v1|" + raw)
+
+
+def _signed(data: dict, priv: Ed25519PrivateKey) -> tuple[bytes, bytes]:
+    raw = json.dumps(data).encode("utf-8")
+    return raw, _sign(raw, priv)
+
+
+# --- load_catalog / validate_catalog: valid round trip ------------------- #
+def test_load_catalog_valid_round_trip():
+    data = _minimal_catalog()
+    raw = json.dumps(data).encode("utf-8")
+    loaded = c.load_catalog(raw)
+    assert loaded == data
+    assert c.validate_catalog(loaded) == []
+    assert c.catalog_version(loaded) == 1
+
+
+def test_load_catalog_accepts_str_too():
+    data = _minimal_catalog()
+    text = json.dumps(data)
+    assert c.load_catalog(text) == data
+
+
+def test_load_catalog_malformed_raises_json_decode_error():
+    # load_catalog is strict json.loads ONLY -- it must never silently
+    # "repair" malformed bytes into something that parses.
+    with pytest.raises(json.JSONDecodeError):
+        c.load_catalog(b"{not valid json")
+
+
+def test_shipped_catalog_json_passes_validation():
+    """Regression test: the real data/catalog.json bundled with the app."""
+    root = Path(c.__file__).resolve().parent
+    raw = (root / "data" / "catalog.json").read_bytes()
+    data = c.load_catalog(raw)
+    problems = c.validate_catalog(data)
+    assert problems == [], problems
+    assert c.catalog_version(data) >= 1
+
+
+# --- verify_catalog_signature --------------------------------------------- #
+def test_verify_catalog_signature_valid():
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    sig = _sign(raw, priv)
+    assert c.verify_catalog_signature(raw, sig, [pub]) is True
+
+
+def test_verify_catalog_signature_tampered_byte_fails():
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    sig = _sign(raw, priv)
+    tampered = bytearray(raw)
+    tampered[0] ^= 0xFF  # flip exactly one byte
+    assert c.verify_catalog_signature(bytes(tampered), sig, [pub]) is False
+
+
+def test_verify_catalog_signature_wrong_key_fails():
+    priv = Ed25519PrivateKey.generate()
+    other_pub = Ed25519PrivateKey.generate().public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    sig = _sign(raw, priv)
+    assert c.verify_catalog_signature(raw, sig, [other_pub]) is False
+
+
+def test_verify_catalog_signature_matches_any_key_in_list():
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    other_pub = Ed25519PrivateKey.generate().public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    sig = _sign(raw, priv)
+    # signing key is second in the list -- rotation support
+    assert c.verify_catalog_signature(raw, sig, [other_pub, pub]) is True
+
+
+def test_verify_catalog_signature_garbage_sig_fails():
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    assert c.verify_catalog_signature(raw, b"not-a-real-signature", [pub]) is False
+    assert c.verify_catalog_signature(raw, b"", [pub]) is False
+
+
+def test_verify_catalog_signature_missing_signature_returns_false():
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    raw = json.dumps(_minimal_catalog()).encode("utf-8")
+    assert c.verify_catalog_signature(raw, None, [pub]) is False
+
+
+def test_verify_catalog_signature_never_raises_on_garbage_inputs():
+    assert c.verify_catalog_signature(b"", b"", []) is False
+    assert c.verify_catalog_signature(b"x", b"y", [b"too-short"]) is False
+    assert c.verify_catalog_signature("not-bytes", b"y", [b"\x00" * 32]) is False
+    assert c.verify_catalog_signature(b"x", b"y", None) is False
+
+
+# --- validate_catalog: per-rule rejections --------------------------------- #
+def test_validate_catalog_rejects_non_dict_root():
+    assert c.validate_catalog(["not", "a", "dict"]) != []
+
+
+def test_validate_catalog_rejects_bad_schema_and_version():
+    data = _minimal_catalog()
+    data["schema"] = 0
+    data["version"] = -1
+    problems = c.validate_catalog(data)
+    assert any("schema" in p for p in problems)
+    assert any("version" in p for p in problems)
+
+
+def test_validate_catalog_basic_requires_config():
+    data = _catalog_with({"config": None})
+    problems = c.validate_catalog(data)
+    assert any("config" in p for p in problems)
+
+
+def test_validate_catalog_link_only_forbids_config():
+    data = _minimal_catalog()
+    data["servers"][0] = {
+        "id": "hosted",
+        "display": "Hosted",
+        "description": "A hosted connector.",
+        "category": "dev",
+        "homepage": "https://example.com/hosted",
+        "official": True,
+        "setup": "link-only",
+        "env_required": {},
+        "docs_url": "https://example.com/hosted/docs",
+        "notes": "",
+        "config": {"command": "npx", "args": ["-y", "should-not-be-here"]},
+    }
+    problems = c.validate_catalog(data)
+    assert any("must not have a 'config'" in p for p in problems)
+
+
+def test_validate_catalog_rejects_disallowed_command():
+    data = _catalog_with({"config": {"command": "bash", "args": ["-c", "echo hi"]}})
+    problems = c.validate_catalog(data)
+    assert any("allowlist" in p for p in problems)
+
+
+def test_validate_catalog_rejects_node_eval_flag():
+    data = _catalog_with({"config": {"command": "node", "args": ["-e", "require('fs')"]}})
+    problems = c.validate_catalog(data)
+    assert any("-e/--eval/-c" in p for p in problems)
+
+
+def test_validate_catalog_rejects_python_c_flag():
+    data = _catalog_with({"config": {"command": "python3", "args": ["-c", "import os"]}})
+    problems = c.validate_catalog(data)
+    assert any("-e/--eval/-c" in p for p in problems)
+
+
+def test_validate_catalog_rejects_docker_privileged():
+    data = _catalog_with(
+        {"config": {"command": "docker", "args": ["run", "--privileged", "some/image"]}}
+    )
+    problems = c.validate_catalog(data)
+    assert any("--privileged" in p for p in problems)
+
+
+def test_validate_catalog_rejects_docker_root_volume_mount():
+    data = _catalog_with(
+        {"config": {"command": "docker", "args": ["run", "-v", "/:/host", "some/image"]}}
+    )
+    problems = c.validate_catalog(data)
+    assert any("mounts" in p for p in problems)
+
+
+def test_validate_catalog_rejects_docker_home_volume_mount():
+    data = _catalog_with(
+        {"config": {"command": "docker", "args": ["run", "--volume=$HOME:/host", "some/image"]}}
+    )
+    problems = c.validate_catalog(data)
+    assert any("mounts" in p for p in problems)
+
+
+def test_validate_catalog_rejects_nonempty_env_required():
+    data = _catalog_with({"env_required": {"API_TOKEN": "sk-shouldnotbehere"}})
+    problems = c.validate_catalog(data)
+    assert any("env_required" in p for p in problems)
+
+
+def test_validate_catalog_rejects_secret_looking_arg():
+    data = _catalog_with(
+        {"config": {"command": "npx", "args": ["-y", "widget-mcp", "--api-key=sk-abcdef123"]}}
+    )
+    problems = c.validate_catalog(data)
+    assert any("secret-looking" in p for p in problems)
+
+
+def test_validate_catalog_rejects_token_prefix_positional_arg():
+    data = _catalog_with(
+        {"config": {"command": "npx", "args": ["-y", "widget-mcp", "ghp_abcdef123456"]}}
+    )
+    problems = c.validate_catalog(data)
+    assert any("secret-looking" in p for p in problems)
+
+
+def test_validate_catalog_rejects_http_url():
+    data = _catalog_with({"homepage": "http://example.com/widget"})
+    problems = c.validate_catalog(data)
+    assert any("homepage" in p for p in problems)
+
+
+def test_validate_catalog_rejects_file_url():
+    data = _catalog_with({"docs_url": "file:///etc/passwd"})
+    problems = c.validate_catalog(data)
+    assert any("docs_url" in p for p in problems)
+
+
+def test_validate_catalog_rejects_non_ascii_id():
+    data = _catalog_with({"id": "wídget"})
+    problems = c.validate_catalog(data)
+    assert any("ASCII" in p for p in problems)
+
+
+def test_validate_catalog_rejects_non_ascii_command():
+    data = _catalog_with({"config": {"command": "npxé", "args": ["-y", "widget-mcp"]}})
+    problems = c.validate_catalog(data)
+    assert any("ASCII" in p for p in problems)
+
+
+def test_validate_catalog_rejects_non_ascii_arg():
+    data = _catalog_with({"config": {"command": "npx", "args": ["-y", "wídget-mcp"]}})
+    problems = c.validate_catalog(data)
+    assert any("non-ASCII" in p for p in problems)
+
+
+def test_validate_catalog_rejects_duplicate_ids():
+    data = _minimal_catalog()
+    data["servers"].append(dict(data["servers"][0]))
+    problems = c.validate_catalog(data)
+    assert any("duplicate id" in p for p in problems)
+
+
+# --- validate_catalog: config.env (issue #68 finding 2) -------------------- #
+def test_validate_catalog_rejects_each_denied_env_key():
+    for key in sorted(c.CATALOG_DENIED_ENV_KEYS):
+        data = _catalog_with(
+            {"config": {"command": "npx", "args": ["-y", "widget-mcp@1.0.0"], "env": {key: ""}}}
+        )
+        problems = c.validate_catalog(data)
+        assert any("deny-list" in p for p in problems), (key, problems)
+
+
+def test_validate_catalog_rejects_denied_env_key_case_insensitively():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"node_options": ""},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert any("deny-list" in p for p in problems)
+
+
+def test_validate_catalog_rejects_nonempty_nonplaceholder_env_value():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"FOO_URL": "https://example.com"},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert any("empty string or a single <PLACEHOLDER>" in p for p in problems)
+
+
+def test_validate_catalog_accepts_empty_and_placeholder_env_values():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"FOO": "", "BAR_URL": "<BAR_URL>"},
+            }
+        }
+    )
+    assert c.validate_catalog(data) == []
+
+
+def test_validate_catalog_rejects_non_ascii_env_key():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"FÖO": ""},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert any("config.env key" in p and "ASCII" in p for p in problems)
+
+
+def test_validate_catalog_rejects_non_ascii_env_value():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"FOO": "<Bäd>"},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert any("config.env value" in p and "ASCII" in p for p in problems)
+
+
+def test_validate_catalog_rejects_secret_looking_env_value():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"SOME_TOKEN": "ghp_abcdef1234567890"},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert any("real secret value" in p for p in problems)
+
+
+def test_validate_catalog_rejects_node_options_env_walking_past_allowlist():
+    # The exact reproduction from issue #68 finding 2: an allowlisted
+    # `npx` command carrying NODE_OPTIONS in env, which previously passed
+    # validation and would have flowed straight into the executed
+    # subprocess via catalog_entry_to_paste_json().
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "npx",
+                "args": ["-y", "widget-mcp@1.0.0"],
+                "env": {"NODE_OPTIONS": "--require /tmp/payload.js"},
+            }
+        }
+    )
+    problems = c.validate_catalog(data)
+    assert problems != []
+
+
+# --- validate_catalog: version pinning (issue #68 finding 3) --------------- #
+def test_validate_catalog_rejects_unpinned_npx_package():
+    data = _catalog_with({"config": {"command": "npx", "args": ["-y", "widget-mcp"]}})
+    problems = c.validate_catalog(data)
+    assert any("not version-pinned" in p for p in problems)
+
+
+def test_validate_catalog_rejects_unpinned_scoped_npx_package():
+    data = _catalog_with({"config": {"command": "npx", "args": ["-y", "@scope/pkg"]}})
+    problems = c.validate_catalog(data)
+    assert any("not version-pinned" in p for p in problems)
+
+
+def test_validate_catalog_accepts_pinned_scoped_npx_package():
+    data = _catalog_with({"config": {"command": "npx", "args": ["-y", "@scope/pkg@1.2.3"]}})
+    assert c.validate_catalog(data) == []
+
+
+def test_validate_catalog_rejects_unpinned_uvx_package():
+    data = _catalog_with({"config": {"command": "uvx", "args": ["some-tool"]}})
+    problems = c.validate_catalog(data)
+    assert any("not version-pinned" in p for p in problems)
+
+
+def test_validate_catalog_accepts_uvx_at_version_pin():
+    data = _catalog_with({"config": {"command": "uvx", "args": ["some-tool@1.0.0"]}})
+    assert c.validate_catalog(data) == []
+
+
+def test_validate_catalog_accepts_uvx_double_equals_pin():
+    data = _catalog_with({"config": {"command": "uvx", "args": ["some-tool==1.0.0"]}})
+    assert c.validate_catalog(data) == []
+
+
+def test_validate_catalog_rejects_docker_latest_tag():
+    data = _catalog_with({"config": {"command": "docker", "args": ["run", "some/image:latest"]}})
+    problems = c.validate_catalog(data)
+    assert any("'latest'" in p for p in problems)
+
+
+def test_validate_catalog_rejects_docker_untagged_image():
+    data = _catalog_with({"config": {"command": "docker", "args": ["run", "some/image"]}})
+    problems = c.validate_catalog(data)
+    assert any("no explicit tag" in p for p in problems)
+
+
+def test_validate_catalog_accepts_pinned_docker_image_with_flags():
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "docker",
+                "args": ["run", "-i", "--rm", "-e", "SOME_TOKEN", "some/image:1.2.3"],
+            }
+        }
+    )
+    assert c.validate_catalog(data) == []
+
+
+def test_validate_catalog_does_not_pin_check_placeholders_flags_or_subcommand():
+    # A pinned uvx spec followed by flags and a <PLACEHOLDER> positional
+    # must not itself get mistaken for an unpinned package.
+    data = _catalog_with(
+        {
+            "config": {
+                "command": "uvx",
+                "args": ["mcp-server-git@2026.7.10", "--repository", "<REPO_PATH>"],
+            }
+        }
+    )
+    assert c.validate_catalog(data) == []
+
+
+# --- validate_catalog: id constraint (issue #68 finding 7) ----------------- #
+def test_validate_catalog_rejects_id_with_markup():
+    data = _catalog_with({"id": "<b>Verified</b>"})
+    problems = c.validate_catalog(data)
+    assert any("must match" in p for p in problems)
+
+
+def test_validate_catalog_rejects_id_with_uppercase():
+    data = _catalog_with({"id": "Widget"})
+    problems = c.validate_catalog(data)
+    assert any("must match" in p for p in problems)
+
+
+def test_validate_catalog_rejects_id_starting_with_dash():
+    data = _catalog_with({"id": "-widget"})
+    problems = c.validate_catalog(data)
+    assert any("must match" in p for p in problems)
+
+
+def test_validate_catalog_accepts_valid_slug_id():
+    data = _catalog_with({"id": "widget-2.thing-ok"})
+    assert c.validate_catalog(data) == []
+
+
+# --- resolve_catalog -------------------------------------------------------- #
+def test_resolve_catalog_nothing_available_returns_empty_dict():
+    assert c.resolve_catalog(None, None, None) == {}
+
+
+def test_resolve_catalog_prefers_highest_verified_version(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    bundled = _signed(_minimal_catalog(version=1), priv)
+    cached = _signed(_minimal_catalog(version=2), priv)
+    remote = _signed(_minimal_catalog(version=3), priv)
+
+    result = c.resolve_catalog(bundled, cached, remote)
+    assert c.catalog_version(result) == 3
+
+
+def test_resolve_catalog_rejects_unsigned_bundled_catalog(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    # Bundled claims a very high version but is NOT signed by a trusted key
+    # -- it must get no implicit trust just for being the local copy.
+    malicious_raw = json.dumps(_minimal_catalog(version=100)).encode("utf-8")
+    bundled = (malicious_raw, b"totally-not-a-signature")
+
+    remote = _signed(_minimal_catalog(version=3), priv)
+
+    result = c.resolve_catalog(bundled, None, remote)
+    assert c.catalog_version(result) == 3
+
+
+def test_resolve_catalog_rejects_rolled_back_version(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    cached = _signed(_minimal_catalog(version=5), priv)
+    rolled_back_remote = _signed(_minimal_catalog(version=2), priv)
+
+    result = c.resolve_catalog(None, cached, rolled_back_remote)
+    assert c.catalog_version(result) == 5
+
+
+def test_resolve_catalog_rejects_absurd_version_jump(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    # The freeze attempt goes FIRST (as `cached`), the legitimate catalog
+    # SECOND (as `remote`) -- on purpose. Putting the good catalog first
+    # (as an earlier version of this test did) never exercises the
+    # vulnerable path: the old implementation only guarded a candidate
+    # against "the best accepted so far," so whichever candidate was
+    # evaluated FIRST got in unconditionally, uncapped. Ordering the freeze
+    # attempt first is what actually proves the cap holds regardless of
+    # evaluation order.
+    freeze_attempt = _signed(_minimal_catalog(version=999999), priv)
+    good = _signed(_minimal_catalog(version=5), priv)
+
+    result = c.resolve_catalog(None, freeze_attempt, good)
+    assert c.catalog_version(result) == 5
+
+
+def test_resolve_catalog_caps_first_and_only_candidate(monkeypatch):
+    # issue #68 finding 6: with no bundled catalog to anchor against, a
+    # signed catalog claiming an absurd version must still be capped even
+    # when it is the ONLY candidate resolve_catalog() ever sees -- there is
+    # no "best so far" for it to be compared against, so the cap has to
+    # apply unconditionally, not "once something else has already landed."
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    freeze_attempt = _signed(_minimal_catalog(version=999999999), priv)
+
+    result = c.resolve_catalog(None, None, freeze_attempt)
+    assert result == {}
+
+
+def test_resolve_catalog_anchors_cap_to_bundled_not_a_chained_best(monkeypatch):
+    # Anti-freeze must be measured against the BUNDLED version specifically,
+    # not against "whatever the best-so-far happens to be after each
+    # candidate is accepted" -- a chained anchor lets each accepted
+    # candidate ratchet the allowed ceiling upward, so a legitimate
+    # moderate bump (cached) plus a second, much larger jump (remote) can
+    # each individually look "within _CATALOG_MAX_VERSION_JUMP of the
+    # previous one" while remote is nowhere near bundled's version.
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    bundled = _signed(_minimal_catalog(version=2), priv)
+    cached = _signed(_minimal_catalog(version=1000), priv)  # within 1000 of bundled
+    remote = _signed(_minimal_catalog(version=1900), priv)  # within 1000 of cached,
+    #                                                          NOT of bundled
+
+    result = c.resolve_catalog(bundled, cached, remote)
+    assert c.catalog_version(result) == 1000
+
+
+def test_resolve_catalog_prefers_bundled_on_version_tie(monkeypatch):
+    # issue #68 finding 6: on a tie the LAST candidate evaluated used to
+    # win, so remote silently beat bundled at equal version. Bundled --
+    # the copy frozen into the binary -- must win ties.
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    bundled_data = _minimal_catalog(version=5)
+    bundled = _signed(bundled_data, priv)
+
+    remote_data = _minimal_catalog(version=5)
+    remote_data["servers"][0]["display"] = "Remote Impostor"
+    remote = _signed(remote_data, priv)
+
+    result = c.resolve_catalog(bundled, None, remote)
+    assert c.catalog_version(result) == 5
+    assert result["servers"][0]["display"] == "Widget"
+
+
+def test_resolve_catalog_floor_rejects_below_persisted_version(monkeypatch):
+    # `floor` is a pure parameter: the caller (eventually the GUI, from
+    # persisted storage) can pass a previously-accepted version, and
+    # nothing below it may be accepted even with no bundled catalog to
+    # anchor against.
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    stale = _signed(_minimal_catalog(version=3), priv)
+
+    result = c.resolve_catalog(None, None, stale, floor=10)
+    assert result == {}
+
+
+def test_resolve_catalog_malformed_candidate_does_not_raise(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    malformed_raw = b"{not valid json"
+    malformed_sig = _sign(malformed_raw, priv)
+    good = _signed(_minimal_catalog(version=1), priv)
+
+    result = c.resolve_catalog((malformed_raw, malformed_sig), None, good)
+    assert c.catalog_version(result) == 1
+
+
+def test_resolve_catalog_invalid_but_signed_candidate_is_skipped(monkeypatch):
+    priv = Ed25519PrivateKey.generate()
+    pub = priv.public_key().public_bytes_raw()
+    monkeypatch.setattr(c, "CATALOG_PUBKEYS", [pub])
+
+    invalid = _signed(_catalog_with({"config": {"command": "bash", "args": []}}), priv)
+    good = _signed(_minimal_catalog(version=1), priv)
+
+    result = c.resolve_catalog(invalid, None, good)
+    assert c.catalog_version(result) == 1
+
+
+# --- catalog_entry_to_paste_json / config_has_unfilled_placeholders ------- #
+def test_catalog_entry_to_paste_json_basic_shape():
+    entry = _minimal_catalog()["servers"][0]
+    result = c.catalog_entry_to_paste_json(entry)
+    assert result == {"widget": {"command": "npx", "args": ["-y", "widget-mcp@1.0.0"]}}
+
+
+def test_catalog_entry_to_paste_json_includes_env_when_present():
+    # NOTE: catalog_entry_to_paste_json() is a pure shape-converter for an
+    # entry that has ALREADY passed validate_catalog() -- it is correct for
+    # it to carry env through verbatim. The bug (issue #68 finding 2) was
+    # never in this function; it was that validate_catalog() let entries
+    # with dangerous/non-placeholder env values reach this function in the
+    # first place. This test now proves that boundary explicitly: a
+    # validation-legal env value (a <PLACEHOLDER> token) survives the
+    # conversion, and a value validate_catalog() would have rejected is
+    # confirmed rejected before it ever gets here.
+    entry = _minimal_catalog()["servers"][0]
+    entry["config"]["env"] = {"GRAFANA_URL": "<GRAFANA_URL>"}
+    result = c.catalog_entry_to_paste_json(entry)
+    assert result["widget"]["env"] == {"GRAFANA_URL": "<GRAFANA_URL>"}
+
+    catalog = _minimal_catalog()
+    catalog["servers"][0]["config"]["env"] = {"GRAFANA_URL": "<GRAFANA_URL>"}
+    assert c.validate_catalog(catalog) == []
+
+    malicious = _minimal_catalog()
+    malicious["servers"][0]["config"]["env"] = {"NODE_OPTIONS": "--require /tmp/payload.js"}
+    assert c.validate_catalog(malicious) != []
+
+
+def test_config_has_unfilled_placeholders_true_for_token():
+    cfg = {"command": "npx", "args": ["-y", "server", "<ALLOWED_DIR>"]}
+    assert c.config_has_unfilled_placeholders(cfg) is True
+
+
+def test_config_has_unfilled_placeholders_false_after_fill():
+    cfg = {"command": "npx", "args": ["-y", "server", "/Users/me/project"]}
+    assert c.config_has_unfilled_placeholders(cfg) is False
+
+
+def test_config_has_unfilled_placeholders_checks_env_too():
+    cfg = {"command": "uvx", "args": ["mcp-grafana"], "env": {"GRAFANA_URL": "<GRAFANA_URL>"}}
+    assert c.config_has_unfilled_placeholders(cfg) is True