.gitea/workflows/secret-scan.yml

@@ -0,0 +1,32 @@
+name: Secret Scan
+
+# **No `paths:` filter, deliberately.** The credential this exists for lived in
+# `app/src-tauri/src/docker/container.rs`, which `build.yml` would have skipped —
+# that workflow only runs for `container/**`. A scan that can be avoided by
+# touching the wrong directory is not a scan.
+#
+# This is the half of the check that nobody can bypass. The pre-commit hook in
+# `.githooks/` is faster and friendlier, but it is opt-in per clone and
+# `--no-verify` skips it; both are true of every git hook and neither is fixable
+# from inside a repository.
+
+on:
+  push:
+    branches: ["**"]
+  pull_request:
+    branches: ["**"]
+
+jobs:
+  scan:
+    runs-on: ubuntu-latest
+    steps:
+      - name: Checkout
+        uses: actions/checkout@v4
+
+      # The whole tracked tree, not just the diff. Scanning a range is cheaper
+      # but depends on getting the range right across pushes, force-pushes,
+      # merges and PR events — and a wrong range fails *open*. The full scan
+      # takes under half a second on this repository and cannot be evaded by
+      # arranging for the interesting commit to sit outside the window.
+      - name: Scan tracked files for credentials
+        run: sh scripts/scan-secrets.sh --tracked

.githooks/pre-commit

@@ -0,0 +1,14 @@
+#!/bin/sh
+# Refuse a commit that adds something shaped like a live credential.
+#
+# Installed by pointing git at this directory:
+#
+#     git config core.hooksPath .githooks
+#
+# which `npm run hooks` in app/ does for you. It is per-clone — git will not let
+# a repository configure its own hooks path, for the obvious reason that cloning
+# a repo would then be enough to run its code. So this is opt-in on every
+# machine, `--no-verify` skips it, and neither of those is a flaw to fix here:
+# the CI job in `.gitea/workflows/build.yml` is the half nobody can bypass. The
+# hook exists to tell you in one second rather than in five minutes.
+exec "$(git rev-parse --show-toplevel)/scripts/scan-secrets.sh" --staged

CLAUDE.md

@@ -525,6 +525,33 @@ Anthropic and Bedrock deliberately keep Claude Code's own defaults.
   `models/project.rs` for anything that should default to true.
 - Cross-platform paths: Docker socket is `/var/run/docker.sock` on Linux/macOS, `//./pipe/docker_engine` on Windows
 
+## Secrets
+
+**`scripts/scan-secrets.sh` refuses a commit that adds something shaped like a live
+credential.** Enable the hook once per clone with `npm run hooks` (from `app/`), which sets
+`core.hooksPath` to `.githooks`. A repository cannot configure its own hooks path — cloning it
+would then be enough to run its code — so this is opt-in everywhere, and `--no-verify` skips it.
+The `Secret Scan` workflow is the half nobody can bypass; it carries **no `paths:` filter**, on
+purpose, because the incident that prompted all this lived in `app/**` and `build.yml` only runs
+for `container/**`.
+
+Three rules, and the second half of the third is what keeps it usable: vendor-prefixed tokens
+(`ghp_`, `sk-`, `AKIA`, `xox`, …), `BEGIN … PRIVATE KEY` blocks, and an opaque literal assigned to
+a secret-shaped name. That last one needs **both** halves — the identifier must read as a
+credential *and* the whole literal must be hex or base64 with no word structure. Name-proximity
+alone flags `secure::get_project_secret(&id, "aws-secret-access-key")`, which is a keychain key
+name; the literal test is what excludes it. Measured against the tree: 0 false positives, and it
+catches the real incident (`9b2f4fe`) when replayed.
+
+A line ending `pragma: allowlist secret` is skipped. Make a fixture obviously fake before reaching
+for it.
+
+**Why this exists:** `the_custom_env_fingerprint_never_carries_the_value` used the maintainer's
+real Gitea **site-admin** token as its fixture — a test about secrets not escaping, leaking one. It
+survived 92 commits and fourteen days in the public GitHub mirror, past five audit rounds and two
+independent reviews, because every one of them read the code under change and this sat in a test
+nobody had reason to open. Fixtures are never live values; there is no case where they need to be.
+
 ## Testing
 
 Frontend tests use Vitest with jsdom environment and React Testing Library. Setup file at `src/test/setup.ts`. Run a single test file:

app/package.json

@@ -9,7 +9,8 @@
     "preview": "vite preview",
     "tauri": "tauri",
     "test": "vitest run",
-    "test:watch": "vitest"
+    "test:watch": "vitest",
+    "hooks": "git -C .. config core.hooksPath .githooks && echo \"pre-commit secret scan enabled\""
   },
   "dependencies": {
     "@tauri-apps/api": "^2",

app/src-tauri/src/docker/container.rs

@@ -4177,7 +4177,12 @@ mod tests {
         // It goes into `triple-c.custom-env-fingerprint`, which `docker inspect`
         // hands to anything on the host, `docker commit` copies onto the
         // project's snapshot image, and the recreation check logs on a mismatch.
-        let secret = "33da01c1b320644920c20d6b5e0a1c6b3c3451c2";
+        // **Never a real credential.** This literal was the maintainer's actual
+        // Gitea token for fourteen days and ninety-two commits, on a public
+        // mirror — in a test whose whole subject is that secrets do not escape.
+        // A fixture only has to be *a value*; it never has to be a live one, so
+        // there is no version of this that justifies pasting something real.
+        let secret = "not-a-real-token-0000000000000000000000";
         let fp = compute_env_fingerprint(&[EnvVar {
             key: "TEA_TOKEN".to_string(),
             value: secret.to_string(),

scripts/scan-secrets.sh

@@ -0,0 +1,111 @@
+#!/bin/sh
+# Refuse to let a live credential into the repository.
+#
+# Written after one got in: `the_custom_env_fingerprint_never_carries_the_value`
+# used the maintainer's real Gitea site-admin token as its fixture. It survived
+# 92 commits and fourteen days in a public mirror, past five audit rounds and two
+# independent reviews — because every one of those looked at the code under
+# change, and this sat in a test nobody had reason to open. A grep would have
+# caught it on the first day. This is that grep.
+#
+# Usage:
+#   scan-secrets.sh --staged        what `git commit` is about to record (the hook)
+#   scan-secrets.sh --range A..B    every line added between two commits (CI)
+#   scan-secrets.sh --tracked       every tracked file, as it stands now
+#
+# Exit 0 clean, 1 on a finding, 2 on misuse.
+#
+# ## Why it scans *added lines* and not the whole file
+#
+# The repository already contains long opaque strings — 317 literals of 32+
+# characters, almost all of them legitimate. A scanner that failed on those would
+# be turned off within a day, which is the normal way this kind of check dies.
+# Judging only what a commit *adds* keeps the signal where a person can act on it.
+#
+# ## Escape hatch
+#
+# A line carrying `pragma: allowlist secret` is skipped. Deliberately wordy: it
+# should be uncomfortable enough to type that it is read as a claim, and it
+# leaves something greppable behind.
+
+set -eu
+
+MODE="${1:---staged}"
+RANGE="${2:-}"
+
+case "$MODE" in
+  --staged)  ADDED=$(git diff --cached --unified=0 --no-color -- . 2>/dev/null || true) ;;
+  --range)   [ -n "$RANGE" ] || { echo "scan-secrets: --range needs A..B" >&2; exit 2; }
+             ADDED=$(git diff --unified=0 --no-color "$RANGE" -- . 2>/dev/null || true) ;;
+  --tracked)
+             # `grep -Iq .` first: without it a binary blob's bytes reach the
+             # rules below, and GNU grep answers "binary file matches" instead
+             # of the line — so a real finding inside one would be reported as
+             # a sentence nobody can act on, and a stray NUL can end the scan
+             # early. Text files only; binaries are not where source secrets
+             # live, and `--staged` never sees them either (git emits
+             # "Binary files differ", not content).
+             ADDED=$(git ls-files -z \
+               | xargs -0 -I{} sh -c 'grep -Iq . "{}" 2>/dev/null && sed "s/^/+/" "{}" 2>/dev/null' \
+               || true) ;;
+  *)         echo "scan-secrets: unknown mode $MODE" >&2; exit 2 ;;
+esac
+
+# Only added lines; drop diff headers (+++ b/path) so a filename never matches.
+CANDIDATES=$(printf '%s\n' "$ADDED" \
+  | grep '^+' \
+  | grep -v '^+++' \
+  | grep -v 'pragma: allowlist secret' \
+  || true)
+
+[ -n "$CANDIDATES" ] || exit 0
+
+FOUND=0
+report() {
+  FOUND=1
+  printf '\n  %s\n' "$1"
+  printf '%s\n' "$2" | sed 's/^/    /' | cut -c1-160
+}
+
+# --- Rule 1: vendor-issued credentials. Shape alone identifies these, so there
+# --- is no false-positive story to tell and no identifier context needed.
+VENDOR=$(printf '%s\n' "$CANDIDATES" | grep -nE \
+  'gh[pousr]_[A-Za-z0-9]{20,}|github_pat_[A-Za-z0-9_]{20,}|glpat-[A-Za-z0-9_-]{20,}|xox[baprs]-[A-Za-z0-9-]{10,}|sk-[A-Za-z0-9]{32,}|AKIA[0-9A-Z]{16}|ASIA[0-9A-Z]{16}|ya29\.[A-Za-z0-9_-]{20,}|AIza[0-9A-Za-z_-]{35}|npm_[A-Za-z0-9]{36}|dckr_pat_[A-Za-z0-9_-]{20,}' \
+  || true)
+[ -n "$VENDOR" ] && report "A vendor-issued credential (its prefix identifies the provider):" "$VENDOR"
+
+# --- Rule 2: private key material.
+KEYS=$(printf '%s\n' "$CANDIDATES" | grep -nE -- '-----BEGIN [A-Z ]*PRIVATE KEY-----' || true)
+[ -n "$KEYS" ] && report "Private key material:" "$KEYS"
+
+# --- Rule 3: an opaque literal assigned to a secret-shaped name.
+#
+# This is the rule that would have caught the Gitea token — `let secret =
+# "<40 hex>"`. Both halves are required, and that is what keeps it usable:
+# the *name* must read as a credential, and the *whole literal* must be hex or
+# base64 with no word structure. `"aws-secret-access-key"` is a keychain key
+# name sitting right next to the word `secret`, and its hyphens are what keep it
+# out; measured against the tree, name-proximity alone flagged four lines of
+# which three were that shape.
+OPAQUE=$(printf '%s\n' "$CANDIDATES" | grep -niE \
+  '(secret|token|api[_-]?key|apikey|passwo?rd|passwd|credential|auth[_-]?(key|token))[^A-Za-z0-9]{0,12}[:=][^"'"'"']{0,12}["'"'"']([0-9a-fA-F]{32,}|[A-Za-z0-9+/]{40,}={0,2})["'"'"']' \
+  || true)
+[ -n "$OPAQUE" ] && report "An opaque literal assigned to a secret-shaped name:" "$OPAQUE"
+
+if [ "$FOUND" -eq 1 ]; then
+  cat >&2 <<'MSG'
+
+  ────────────────────────────────────────────────────────────────────────
+  Refusing the commit: it adds something shaped like a live credential.
+
+  If it IS live: do not amend and move on. Rotate it first — anything that
+  reaches a branch is on the mirror, and the mirror is public.
+
+  If it is genuinely not a secret — a fixture, a public key id, test data —
+  make the literal obviously fake ("not-a-real-token-0000…"), or append
+  `pragma: allowlist secret` to the line to say so on the record.
+  ────────────────────────────────────────────────────────────────────────
+MSG
+  exit 1
+fi
+exit 0