@coo-quack/sensitive-canary 0.7.0 → 0.8.1

package/README.md

@@ -20,14 +20,16 @@ Claude Code is a powerful development tool, but file reads and command execution
 | Without sensitive-canary | With sensitive-canary |
 |--------------------------|----------------------|
 | `cat .env` → full contents sent to Claude ❌ | Blocked by name before Claude reads it ✅ |
-| Paste `AKIAIOSFODNN7EXAMPLE` in prompt ❌ | Blocked before the API call is made ✅ |
-| Tool result contains user@email.com ❌ | PII detected and blocked ✅ |
+| Paste a live AWS key in a prompt ❌ | Blocked before the API call is made ✅ |
+| `Read customers.csv` full of email addresses ❌ | PII detected before Claude sees the file ✅ |
 | `echo $API_KEY` with live key ❌ | Env var value scanned and blocked ✅ |
+| `cat docker-compose.yml` with `POSTGRES_PASSWORD:` ❌ | Assignment detected in YAML and JSON too ✅ |
 
 - **Two hooks** — `UserPromptSubmit` and `PreToolUse` cover both directions of risk
-- **64 detection rules** — sourced from gitleaks and TruffleHog detector definitions
+- **76 detection rules** — sourced from gitleaks and TruffleHog detector definitions
 - **Checksum validation** — credit cards (Luhn) and national ID numbers (JP My Number, FR NIR, IT Codice Fiscale, DE Steuer-IdNr., ES DNI/NIE, KR RRN/BRN, CN Resident ID)
-- **Context gating** — phone numbers, postal codes, and public IP addresses require a nearby label, reducing false positives on bare digit sequences
+- **Context gating** — the noisiest rules only fire when a label is nearby: non-US/JP phone numbers, ZIP, EU/KR and Chinese postal codes, public IPv4 and IPv6, and the Korean resident and business numbers. US and Japanese phone numbers and Japanese postal codes are matched without a label, since their shapes are specific enough. RFC 1918 private addresses are not matched at all — they are non-routable, they identify nothing outside the network they belong to, and they fill the inventories, manifests and ssh configs this tool is most often pointed at
+- **Not everything that looks like a secret is one** — published test card numbers, RFC 2606 domains (`example.com`), a value that is a variable reference (`PASSWORD: ${VAR}`), an ssh or scp target (`git@github.com`, `deploy@host`, `user@host:path`), and `.env.example` and its siblings are left alone. Each was blocking ordinary work. A template is exempt only when its contents can be read whole. One holding a NUL byte, running past the per-file cut, reached after the call's budget or deadline, or that is not a regular file at all is blocked on its name, since the contents are what the exemption relies on. A template name that exists on no disk is not blocked — there is nothing to read and nothing to leak
 - **Entropy filtering** — reduces false positives on low-entropy values
 - **Local only** — all scanning runs in your terminal; nothing is sent anywhere
 
@@ -56,7 +58,15 @@ Install in two commands from inside a Claude Code session:
 /plugin install sensitive-canary@coo-quack
 ```
 
-Done. The hooks are enabled automatically.
+The hooks are enabled for every session started after this. A session that was
+already running does not pick them up — it reports the plugin as enabled and
+checks nothing — so start a new one.
+
+**Then check that it blocks.** An installation that checks nothing looks exactly
+like one that works, and only exit 2 stops a tool call, so a hook that fails to
+start is silent. Write a file holding `AKIA` followed by `IOSFODNN7EXAMPLE` and
+ask Claude to read it. It should refuse and say why. If it shows you the key, the
+hooks are not running.
 
 > **Keeping up to date:** Third-party marketplaces have auto-update disabled by default. To receive automatic updates, run `/plugin` → **Marketplaces** tab → select the marketplace → **Enable auto-update**. You can also update manually from the same tab. See [Discover and install plugins](https://docs.anthropic.com/en/docs/claude-code/discover-plugins) for details.
 
@@ -85,18 +95,18 @@ Then add to `~/.claude/settings.json`:
         "hooks": [
           {
             "type": "command",
-            "command": "npx tsx $(npm root -g)/@coo-quack/sensitive-canary/src/user-prompt-submit-hook.ts"
+            "command": "node $(npm root -g)/@coo-quack/sensitive-canary/dist/user-prompt-submit-hook.js"
           }
         ]
       }
     ],
     "PreToolUse": [
       {
-        "matcher": "Read|Bash",
+        "matcher": "Read|NotebookRead|Bash|Grep|mcp__.*",
         "hooks": [
           {
             "type": "command",
-            "command": "npx tsx $(npm root -g)/@coo-quack/sensitive-canary/src/pre-tool-use-hook.ts"
+            "command": "node $(npm root -g)/@coo-quack/sensitive-canary/dist/pre-tool-use-hook.js"
           }
         ]
       }
@@ -105,7 +115,7 @@ Then add to `~/.claude/settings.json`:
 }
 ```
 
-> **Note:** Node.js does not support `--experimental-strip-types` for files inside `node_modules`, so `npx tsx` is used instead.
+> **Note:** These point at the compiled JavaScript the package ships. Node refuses to strip types from a `.ts` file inside `node_modules`, and a hook that fails to start exits non-zero without blocking — so an installation wired to `src/` looks installed and checks nothing. The plugin install uses the `.ts` sources, which sit outside `node_modules` and work.
 
 </details>
 
@@ -141,7 +151,7 @@ Then add to `~/.claude/settings.json`:
     ],
     "PreToolUse": [
       {
-        "matcher": "Read|Bash",
+        "matcher": "Read|NotebookRead|Bash|Grep|mcp__.*",
         "hooks": [
           {
             "type": "command",
@@ -165,11 +175,14 @@ Then add to `~/.claude/settings.json`:
 Prompts containing secrets or PII are blocked before being sent.
 
 ```
-> My AWS key is AKIAIOSFODNN7EXAMPLE. Can you review this code?
+> Here is my deploy key:
+> -----BEGIN RSA PRIVATE KEY-----
+> MIIEowIBAAKCAQEAwK3vJ9m5Q8xY2nB4dF6hL0pR7sT1uV3wX5yZ8aC2eG4iK6mO
+> Can you review the config that uses it?
 
-🐤  sensitive-canary: sensitive data detected — blocked
+🐦 sensitive-canary: sensitive data detected — blocked
 
-  [Secret] AWS Access Key ID (aws-access-key): AKIA****MPLE
+  [Secret] PEM Private Key (private-key): ---****KEY
 
 To allow, add a tag to your prompt:
   [allow-secret]  — allow secrets
@@ -179,12 +192,12 @@ To allow, add a tag to your prompt:
 To allow it through, add the suggested tag:
 
 ```
-> [allow-secret] My AWS key is AKIAIOSFODNN7EXAMPLE. Can you review this code?
+> [allow-secret] Here is my deploy key: -----BEGIN RSA PRIVATE KEY-----
 ```
 
 ### .env file blocked
 
-`.env` / `.env.*` files are blocked by filename, regardless of their contents. This name-based block is a secret guard and only applies while the `secret` category is enabled (the default).
+`.env` and its siblings are blocked by filename, before anything is read. Template names — `.env.example`, `.env.sample`, `.env.template`, `.env.dist`, `.env.defaults` — are the exception: they are meant to be committed, so they are read and judged on their contents like any other file. A template that turns out to hold a real credential is still blocked, and one whose contents cannot be read whole falls back to the name. This name-based block is a secret guard and only applies while the `secret` category is enabled (the default).
 
 ```
 > Read .env
@@ -225,7 +238,7 @@ To intentionally bypass a block, include the appropriate tag in your **current p
 | `[allow-pii]` | Skip all PII-category checks |
 | `[allow-all]` | Skip all sensitive-canary checks |
 
-> **Note:** Tags are read from the **current user message only**. Tags in previous messages are ignored — there is no risk of an accidental persistent bypass. Tags are case-insensitive. `[allow-secret]` does not bypass PII blocks (and vice versa). The name-based block on `.env`/`.env.*` files can be bypassed by any of the three allow tags.
+> **Note:** Tags are read from the **current user message only**. Tags in previous messages are ignored — there is no risk of an accidental persistent bypass. Tags are case-insensitive. `[allow-secret]` does not bypass PII blocks (and vice versa). The name-based block on `.env`/`.env.*` files is a secret guard, so `[allow-secret]` and `[allow-all]` lift it and `[allow-pii]` does not.
 
 ---
 
@@ -319,15 +332,16 @@ User rules support the same fields as built-in rules:
 |---|---|---|
 | `requireContext` | boolean | Only fire when a nearby context word is found |
 | `contextWords` | string[] | Words that satisfy the context requirement |
+| `excludeContext` | string[] | Words that, found nearby, say the match is not what the rule is after — the mirror of `contextWords` |
 | `contextWindow` | number | Override the global context window (default: 3 tokens) |
 | `entropyThreshold` | number | Skip matches below this Shannon entropy |
-| `secretGroup` | number | Capture group index containing the secret (default: 0 = full match) |
+| `secretGroup` | number | Capture group holding the secret. Omit for the whole match — writing `0` is not the same as omitting it, see [Detection Rules](https://coo-quack.github.io/sensitive-canary/rules.html) |
 | `validate` | string | Name of a built-in checksum validator (see below) |
-| `flags` | string | Regex flags (default: `"g"`) |
+| `flags` | string | Regex flags. `g` is added if left out; `y` makes a rule match only at the very start of the text |
 
 Available validators (referenced by name in the `validate` field):
 
-`luhn`, `mynumber-jp`, `nir-fr`, `codice-fiscale-it`, `steuer-id-de`, `dni-nie-es`, `rrn-kr`, `brn-kr`, `resident-id-cn`, `public-ipv4`, `public-ipv6`
+`luhn`, `aws-key`, `phone-jp`, `mynumber-jp`, `nir-fr`, `codice-fiscale-it`, `steuer-id-de`, `dni-nie-es`, `rrn-kr`, `brn-kr`, `resident-id-cn`, `public-ipv4`, `public-ipv6`
 
 ### Overriding the context window globally
 
@@ -346,13 +360,26 @@ Invalid rules (bad regex, wrong types, missing required fields) are skipped with
 
 ## Detection rules
 
-### Secrets (39 rules)
+### Secrets (52 rules)
 
 | Rule ID | Description |
 |---|---|
+| `openai-service-key` | OpenAI Service Account / Admin Key (`sk-svcacct-`, `sk-admin-`, `sk-proj-` prefix) |
+| `azure-storage-key` | Azure Storage Account Key (`AccountKey=` + 88-char base64) |
+| `azure-sas-key` | Azure Shared Access Key for Service Bus, Event Hubs and IoT Hub (`SharedAccessKey=` + 44-char base64). Separate from the storage account key, which is 88 characters |
+| `google-oauth-secret` | Google OAuth Client Secret (`GOCSPX-` prefix) |
+| `flyio-token` | Fly.io API Token (`FlyV1 fm2_` prefix) |
+| `databricks-token` | Databricks Personal Access Token (`dapi` + 32 hex) |
+| `vault-token` | HashiCorp Vault Token (`hvs.` / `hvb.` prefix) |
+| `shopify-token` | Shopify Access Token (`shpat_`, `shpss_`, `shpca_`, `shppa_` prefix) |
+| `doppler-token` | Doppler Token (`dp.pt.`, `dp.st.`, … prefix) |
+| `grafana-token` | Grafana Cloud / Service Account Token (`glc_`, `glsa_` prefix) |
+| `notion-token` | Notion Integration Token (`ntn_` prefix) |
 | `aws-access-key` | AWS Access Key ID |
 | `gcp-api-key` | Google Cloud API Key |
 | `private-key` | PEM Private Key (RSA / EC / DSA / PGP / OpenSSH) |
+| `private-key-base64` | PEM private key that has been base64-encoded — how one appears in a kubeconfig, a Kubernetes Secret or a Terraform state, where the `-----BEGIN` header never shows in the text |
+| `url-basic-auth` | Credentials in the userinfo field of an http(s) URL — a git remote, a `.netrc`, a private registry, a `curl` invocation. RFC 3986 deprecates the form for this reason |
 | `github-pat` | GitHub Personal Access Token |
 | `github-fine-grained` | GitHub Fine-Grained Token |
 | `gitlab-pat` | GitLab Personal Access Token |
@@ -390,13 +417,12 @@ Invalid rules (bad regex, wrong types, missing required fields) are skipped with
 | `env-assignment` | `.env`-style secret assignment *(entropy ≥ 3.0)* |
 | `connection-string` | Database connection string with embedded credentials |
 
-### PII (25 rules)
+### PII (24 rules)
 
 | Rule ID | Description | Validation |
 |---|---|---|
 | `pii-email` | Email address | — |
 | `pii-credit-card` | Credit card number | Luhn check |
-| `pii-ipv4` | IPv4 address (RFC 1918 private ranges only) | — |
 | `pii-ssn` | US Social Security Number | Invalid prefix exclusion |
 | `pii-mynumber-jp` | Japanese Individual Number (My Number) | Checksum (weighted mod 11) |
 | `pii-nir-fr` | French NIR / Social Security Number | Check key (mod 97) |
@@ -449,30 +475,87 @@ When blocked, the terminal shows what was detected and how to bypass it.
 
 ### ② PreToolUse hook
 
-Runs just before Claude calls the `Read` or `Bash` tool.
+Runs just before Claude calls the `Read`, `Bash` or `Grep` tool, or any MCP tool.
 
 ```
-Claude calls Read / Bash tool
+Claude calls Read / Bash / Grep / MCP tool
       ↓
 PreToolUse hook
       ↓
       ── Read tool ─────────────────────────────────────────────────────
       │  1. filename is .env / .env.* → blocked (secret category only)
       │  2. file contents contain secret / PII → blocked
-      └─ Bash tool ─────────────────────────────────────────────────────
-         1. env var values referenced in the command contain secret / PII → blocked
-         2. command string itself contains secret / PII (e.g. echo AKIA...) → blocked
-         3. cat / head / tail / etc. targeting a file → file contents scanned
+      │
+      ├─ Bash tool ──────────────────────────────────────────────────────
+      │  1. env var values referenced in the command contain secret / PII → blocked
+      │  2. a bare env / printenv would print the whole environment → every
+      │     variable is scanned
+      │  3. command string itself contains secret / PII (e.g. echo AKIA...) → blocked
+      │  4. the command is located past any wrapper (sudo, env VAR=1, timeout,
+      │     nice, xargs) and any leading VAR=value assignment
+      │  5. inline scripts (-c, -e, -pe) are parsed and scanned
+      │  6. file paths from input redirections, command substitutions and chained
+      │     commands are extracted and scanned
+      │  7. printing commands (cat, head, tail, sed, awk, grep, rg, cut, sort,
+      │     base64, xxd, strings, diff, comm, dd, and git subcommands) targeting
+      │     a named file → file contents scanned
+      │
+      └─ every other tool, Grep and mcp__* included ─────────────────────
+         1. input fields naming an existing file are scanned for
+            secret / PII → blocked
 ```
 
-When blocked, Claude receives a JSON response explaining the reason and is prompted to tell the user.
+A value is scanned when either its field name says path or the value itself is shaped like one.
+
+The field names are `path`, `paths`, `file`, `files`, `filepath`, `filename`, `filenames`, `absolutepath`, `notebookpath` and `sourcepath`, compared with separators and case removed — so `file_path`, `filePath` and `filepath` are one name. Beyond those, any value containing a `/` is treated as a path whatever its field is called, which is what covers a tool carrying its path under `target`, `document` or `uri`.
+
+The `/` is what separates a path from a word, and it is there so that a search pattern is not read as a path: `{ "pattern": ".env" }` is a search for the text `.env`, not a read of the file, and `.env` exists in most checkouts. The cost is that a bare filename under an unlisted field name is still missed.
+
+Values are found up to four levels down and inside arrays, both of strings and of objects, so `{ "path": "…" }`, `{ "paths": ["…"] }`, `{ "args": ["/abs/…"] }` and `{ "args": { "paths": [{ "path": "…" }] } }` are all covered. A field naming a directory is left alone.
+
+Which tools reach the hook at all is the matcher's business, and the default (`Read|NotebookRead|Bash|Grep|mcp__.*`) sends it `Read`, `NotebookRead`, `Bash`, `Grep` and every MCP tool. Widen the matcher and the same field search applies to whatever else arrives.
+
+Commands that only measure a file (`wc`, `cksum`, `sha256sum`) are not treated as reads, whether the file is named or fed in over `<`: they print counts and digests, never the bytes. Neither are the tools that surface no file contents — `Write`, `Edit`, `MultiEdit`, `NotebookEdit`, `TodoWrite`, `Glob`, `WebFetch`, `WebSearch`, `ExitPlanMode`, `AskUserQuestion` — nor any tool whose name leads with a write verb, such as `mcp__fs__write_file` or `createPage`.
+
+Neither is a command that sends its result back to the file it was handed. `sed -i`, `perl -i` and `ruby -i` (bundled forms such as `perl -pi -e` and `perl -lpi` included) edit in place and write nothing to stdout. A bundle is read one letter at a time, continuing only past switches that command is known to accept without a value — so `sed -Ei` and `perl -lpi` are in-place edits, while `perl -Ilib -pe` and `perl -MList::Util -pe` are reads. A letter the list does not know stops the reading and the file is scanned, which is the safe way to be wrong. `git log <file>` is not a read either — it prints who changed the file and when — unless a patch is asked for with `-p`, `-u`, `--patch`, `-U<n>`, `--unified=<n>`, one of the merge-diff forms (`-c`, `-m`, `--cc`, `--diff-merges`), or `-L`, which prints the lines of one named file.
+
+When blocked, the hook exits 2, which stops the tool call, and writes the reason to stderr, which is where Claude reads it from. The reason names what was detected and which allow tag lifts the block, and asks Claude to pass that on to the user.
 The terminal also receives a direct message (via `/dev/tty`).
 
+### Known Limitations
+
+- **Heredoc bodies** — a heredoc body is treated as text, not as commands, so `cat > deploy.sh <<'EOF'` writing a script that mentions `.env` is not itself a read. The trade-off is that a heredoc which *feeds* commands to another shell (`ssh host <<'EOF'` with a `cat /etc/secrets` in the body) is not inspected either.
+- **A tool that runs a command is read for the command, by field name** — `command`, `commands`, `cmd`, `script`, `code`, `commandline` and `shellcommand`, each read with punctuation and case ignored, so `command_line`, `command-line`, `commandLine` and `command.line` are the same name. A shell-running MCP server that names the field something else hands its command past unread.
+- **Only the first and last 1 MiB of a file are scanned** — a file larger than the cut is read at both ends rather than to its end, because a hook that does not return is killed by the PreToolUse timeout, and a killed hook does not block the call. The cut is in bytes, so a file of multi-byte characters gives up sooner in characters. What is missed is the middle of a file larger than both windows, and a secret straddling either edge, since the cut lands mid-match; the 64 KB transcript tail read makes the same trade. What the cut does *not* bound is the work done on what it read: that is a property of each rule's pattern, and `docs/rules.md` covers why three of them carry length bounds.
+- **A write-named tool that also returns contents** — the exemption reads a tool's name, and assumes a name led by a write verb means the tool surfaces no file contents. `update` and `copy` are where those two things come apart: `mcp__*__update_file` and `mcp__*__copy_file` open a file to do their work, and one that returned the result would not be scanned. Scanning them instead would block writing to a file that already holds a secret, which is not a leak, so the exemption stays as it is.
+- **A bare filename under an unlisted field name** — a value is treated as a path when its field name says so or when it contains a `/`. A tool passing `{ "target": "secrets.txt" }` satisfies neither, so it is not scanned. Requiring the `/` is deliberate: without it, a search for the text `.env` would be blocked as though the file had been read.
+- **git history references** — `git show HEAD:.env` and similar references to objects in git history (not on disk) are not scanned, since the object does not exist as a file path.
+- **Unlisted commands** — the set of commands known to print file contents is a list, not an analysis of the command. A printing command that is not on the list is not caught.
+- **A template holding a real credential is blocked** — `.env.example` and its siblings are exempt from the name guard, not from the scan. Placeholders (`your-token-here`, `REPLACE_ME`, `changeme`, `<token>`, `postgres://user:password@localhost/db`) are recognised and left alone, but a template committed with a live key is blocked like any other file, through printing commands (`grep KEY .env.example`) as much as through `Read`. Use `[allow-secret]` if that is deliberate.
+- **A binary swept up by a directory being named** — the files under a directory are scanned because the directory was named, and one whose first four kilobytes read as neither text nor UTF-16 is skipped rather than ground through every rule. A file named outright is scanned whatever its bytes look like, and so is an `.env` in a swept directory: the name decides that one when the contents cannot.
+- **Anything that is not a regular file** — a FIFO, a process substitution (`/dev/fd/63`) and `/dev/stdin` are not read, so `cat` of one is not scanned. Reading them can never reach the end of the file: `cat /dev/zero` held the hook open until Claude Code's PreToolUse timeout killed it, and a killed hook does not block the call. Not scanning them is the lesser of the two, since a hang lets the call through as well. A directory is not read either, but it is not left alone: the files directly under it are scanned, one level and no further.
+- **A search that names no path is judged on names alone** — `rg pattern`, `grep -r pattern` and `Grep {pattern}` with no `path` all print from the working directory, so that directory is checked for an `.env` and its siblings. Its other files are not read. A directory the user named is one they asked about and its contents are scanned; a directory only implied by a search is every repository anyone works in, and reading those stopped a plain `rg TODO` in a third of the checkouts it was measured against.
+- **`~user/…` is not expanded** — `~` and `~/…` are resolved to the home directory, but the form naming another user needs the password database, and guessing would name the wrong file.
+- **A file in an encoding neither reading recovers** — every run of text between NUL bytes is scanned, and UTF-16 is decoded first: by its byte-order mark, or by NULs falling on one side of each pair through the first sixteen kilobytes. Without a mark that verdict is a guess, so both readings are scanned and a wrong guess hides nothing. What is left out is an encoding that is neither: a file in Shift_JIS or GBK is read as the bytes it is, and a credential written in one of those character sets is not found. An ASCII credential inside such a file still is.
+- **A shell construct that names the file only at run time** — `for f in secrets; do cat "$f"; done` and `find . -name secrets -exec cat {} +` both name the file in the command line, but the hook classifies the command it can see, and in these the reading command is `cat` reached through a loop or through `find`'s own argument list.
+- **At most 64 MiB is read across one tool call** — the per-file cut bounds one file; this bounds the call. A glob naming three hundred large files took half a minute, which is long enough for the PreToolUse timeout to kill the hook, and a killed hook does not block. Files past the budget are not scanned, so naming enough large files before the one that matters is a way past the scan.
+- **A relative path is resolved against the directory Claude Code reports** — and against a literal `cd` at the start of the same command. A `cd` later in the line, one inside a subshell, and one whose argument is a variable, a glob or `-` are all left alone, because where they land cannot be worked out here. A directory changed some other way is the same case.
+- **`**` reaches one level, not every level** — a pattern crossing directories is expanded as a single `*`, because expanding it properly walked a whole tree until the hook was killed. `cat **/secrets` sees `*/secrets`.
+- **One tool call stops reading after five seconds** — whatever it has read by then is what was scanned. A byte budget bounds the reading; this bounds the walking as well, and both are ways past the scan for anyone willing to name enough files first.
+- **A glob is expanded by the hook, not by the shell** — `cat *.env` is expanded here to decide what to scan, a moment before the shell expands it and against the hook's own working directory. A file created in between is missed, and at most 256 matches of one pattern are scanned.
+- **Paths held in shell variables** — a path is only scanned when it appears literally in the command. `f=.env; cat "$f"` resolves at run time, after the hook has already decided.
+- **Paths arriving over a pipe** — `find . -name '.env' | xargs cat` names no file the hook can see.
+- **Programs that read files themselves** — `python script.py` is not scanned, because running a script does not print its source; whatever the script opens at run time is beyond the hook's reach.
+- **A flag's separate value is collected as a path** — on a printing command, only a few flags are known to take a value, so every other flag's value becomes a path candidate: the `5` in `head -n 5 f` and in `cut -c 5 f`. A pattern-first command is different — `grep -A 5 f` spends the `5` as the pattern instead, and only `f` is collected. Harmless in practice, since only paths that exist as regular files are read — a file named `5` in the working directory would be scanned, and nothing else is.
+- **A command a wrapper hands off to may be mistaken for its argument, or the reverse** — the search past `sudo`, `timeout` and the others takes the first name it can classify, because a wrapper flag's value (`sudo -u root cat f`) cannot be told apart from a command name. So an unclassified command's arguments are searched too: `sudo mycmd cat f` resolves to `cat` and scans `f`. `echo`, `printf`, `true`, `false` and `:` are known to print their arguments rather than open them, and stop the search; any other unclassified name does not.
+- **Inline program text is followed four levels deep** — each `-c` / `-e` script inside another costs one level, so a read buried five interpreters down is not reached. Nested command substitutions are not bounded this way.
+- **Best effort only** — detection is not exhaustive. Arbitrary shell metacharacters, eval chains, and complex expansions may not be fully tracked.
+
 ---
 
 ## Allow Tags (detailed)
 
-Allow tags filter the scan results — the scan still runs. The `.env`/`.env.*` name block is the only exception: when an allow tag is present, the file is passed through immediately without scanning.
+Allow tags filter the scan results — the scan still runs, including for a `.env` file whose name guard a tag has lifted. Lifting the name guard is not the same as skipping the check: `[allow-secret]` on a `.env` holding an email address still blocks on the address.
 
 ### Mask tags
 
@@ -481,14 +564,14 @@ Allow tags filter the scan results — the scan still runs. The `.env`/`.env.*`
 If you include a mask tag, sensitive-canary will explain this and list what was detected:
 
 ```
-> [mask-secret] My key is AKIAIOSFODNN7EXAMPLE, can you review this?
+> [mask-secret] My deploy key is -----BEGIN RSA PRIVATE KEY----- , can you review this?
 
 🐦 sensitive-canary: prompt masking is not supported
 
   [mask-secret] cannot mask prompt content.
   The following sensitive data was detected:
 
-  [Secret] AWS Access Key ID (aws-access-key): AKIA****MPLE
+  [Secret] PEM Private Key (private-key): ---****KEY
 
   Please choose one of the following:
 
@@ -500,13 +583,19 @@ If you include a mask tag, sensitive-canary will explain this and list what was
 
 ### Allow + Mask tag priority
 
-When both `[allow-*]` and `[mask-*]` tags appear in the same prompt, **the tag that appears first wins** for each category (`secret`, `pii`). `[allow-all]` and `[mask-all]` resolve both categories at once.
+When more than one tag appears, **the last one wins**. It replaces the earlier ones entirely rather than combining with them, so changing your mind mid-message works the way it reads.
 
-| Example | Result |
-|---------|--------|
-| `[allow-secret] [mask-secret] …` | secret allowed |
-| `[mask-secret] [allow-secret] …` | masking not supported error |
-| `[allow-secret] [mask-pii] …` | secret allowed, PII mask error |
+| Example | secret | pii |
+|---------|--------|-----|
+| `[allow-all] … [allow-secret]` | allow | blocked |
+| `[allow-secret] … [allow-all]` | allow | allow |
+| `[allow-secret] … [mask-secret]` | mask (unsupported) | blocked |
+| `[mask-secret] … [allow-secret]` | allow | blocked |
+| `[allow-secret] … [allow-pii]` | blocked | allow |
+
+The last line is the one to watch: two tags do not add up. Narrowing from `[allow-all]` to `[allow-secret]` really does put PII back under guard, which is the point — but so does writing `[allow-secret] [allow-pii]` and expecting both. **`[allow-all]` is how you ask for both.**
+
+A tag counts wherever it appears in the message, mid-sentence included. What does not count is a tag inside a fenced code block, inside one of the elements Claude Code writes around command output, or in a line the runtime wrote rather than you — a compaction summary, a skill body, or a background task reporting back. Those are quoting, not asking.
 
 ---
 
@@ -523,6 +612,10 @@ src/
   lib/
     inspector.ts               allow tag parsing, message scanning
     rules.ts                   secret and PII detection rule definitions
+    default-config.json        the rules themselves, as data
+    shell.ts                   shell syntax: tokens, quoting, heredocs, substitutions
+    bash-commands.ts           what each command does with the files it is given
+    tool-inputs.ts             which input fields of a tool name a file
 ```
 
 ---
@@ -530,12 +623,16 @@ src/
 ## Development
 
 ```bash
-npm install        # install dependencies
+pnpm install        # install dependencies
 
-npm test           # run tests
-npm run test:watch # run tests in watch mode
-npm run typecheck  # type check (tsc)
-npm run lint       # lint with Biome (no changes)
-npm run fix        # lint + auto-fix with Biome
-npm run ci         # typecheck + lint + tests (for CI)
+pnpm test           # run tests
+pnpm run test:watch # run tests in watch mode
+pnpm run typecheck  # type check (tsc)
+pnpm run lint       # lint with Biome (no changes)
+pnpm run fix        # lint + auto-fix with Biome
+pnpm run ci         # typecheck + lint + tests (for CI)
 ```
+
+The lockfile is pnpm's, and every CI job installs with pnpm, so `npm install`
+here ignores it, writes a second lockfile, and resolves a different tree from the
+one that is tested. `CONTRIBUTING.md` has the rest of the workflow.