Ban PR Citations In Rule Text

corpus/CLAUDE.learned.md

@@ -19,7 +19,7 @@ Engine-only install drops this file; reflect Accepted global rules land here.
 - When a fix changes which identity, binary, or database a process uses, verify the one the *service* resolves, not the one your shell resolves. Same session: `which invoker-cli` from an ssh login returned a 0.1.0 system install under `/usr/lib/node_modules`, while the owner process's own `PATH` resolved a different 0.1.1 binary under `~/.local/npm-global/bin`; and a worker's inventory query shelled a checkout script that opened a database holding a week-old snapshot (60 workflows, newest 8 days old) instead of the live owner's (18 workflows, newest that day).
 - Absence of a field in a projection (CLI, MCP, or API output) is not proof of absent state. Say "not projected" and find the emitter before retrying, resubmitting, or switching agents. Seen once: a task sat in `needs_input` with its reason stored by the owner but omitted by every headless query; the session guessed three causes (usage limit, agent switch, dirty branch) and resubmitted twice before reading the emitter, which then took four minutes.
 - When delegating a file-finding task to a subagent and two files could plausibly hold the same answer (a duplicate, a moved file, a same-named symbol in two packages), tell the subagent to state whether each file:line claim is "read-confirmed" (it opened the actual reference/import and traced it) or "name-matched" (it assumes the file is the one in use because the name/path looked right). A subagent that reasons by name-proximity instead of tracing the real reference can hand back a confident wrong file — a judgment call about how the subagent qualifies its own confidence, not something a mechanical check can catch. Seen once: a subagent named a dead, unused duplicate file as the source of a bug; the citation was relayed as fact for three turns before a direct Read caught it.
-- A `file:line` citation, mine or a subagent's, also names the ref it was read at: working tree, `HEAD`, `origin/<base>`, or the installed bundle. A working-tree read in a checkout with untracked or modified files under the cited path is name-matched, not read-confirmed, until the same line is shown at the ref the change will actually run on. Subagent prompts that ask for read-confirmed vs name-matched must also ask for the ref. Seen repeatedly: a fix was planned against `scripts/e2e-regression-watch.mjs:330` read on a dirty feature branch after `origin/master` had already removed it (#11593); a launch-time freshness gate was misdiagnosed from an untracked copy of `task-specification-preflight.ts`; three Explore subagents reported branch-local lines as read-confirmed. Third instance after the memory `verify-bug-findings-against-master`.
+- A `file:line` citation, mine or a subagent's, also names the ref it was read at: working tree, `HEAD`, `origin/<base>`, or the installed bundle. A working-tree read in a checkout with untracked or modified files under the cited path is name-matched, not read-confirmed, until the same line is shown at the ref the change will actually run on. Subagent prompts that ask for read-confirmed vs name-matched must also ask for the ref. Seen repeatedly: a fix was planned against `scripts/e2e-regression-watch.mjs:330` read on a dirty feature branch after `origin/master` had already removed it (Invoker #11593); a launch-time freshness gate was misdiagnosed from an untracked copy of `task-specification-preflight.ts`; three Explore subagents reported branch-local lines as read-confirmed. Third instance after the memory `verify-bug-findings-against-master`.
 - When a pipeline switches from test/synthetic inputs to my real inputs, re-derive or explicitly re-validate every artifact built under the old conditions — voice clones, cached device lists, presigned URLs, browser sessions that predate a driver install. State which artifacts were rebuilt and which were kept. Seen once: a voice clone built from a bench test-mp3 silently became the live demo voice ("WHY IS MY VOICE ROBOTIC?"), and a Chrome that predated the BlackHole driver install couldn't enumerate it until fully relaunched.
 
 # Session hygiene (apply everywhere, every project)

engine/hooks/skill-usage-log/tests/test_hooks.py

@@ -64,7 +64,9 @@ def test_disabled_flag_writes_nothing(self):
                 state_dir=tmp,
             )
             self.assertEqual(read_lines(tmp), [])
-            self.assertFalse(os.path.exists(claude_pretooluse_log.log_path()))
+            self.assertFalse(
+                os.path.exists(os.path.join(tmp, claude_pretooluse_log.LOG_FILE_NAME))
+            )
 
     def test_missing_skill_name_writes_nothing(self):
         with tempfile.TemporaryDirectory() as tmp:

product/skills/ship-a-detector/SKILL.md

@@ -22,7 +22,7 @@ Measured on this repo's own history: 65 of 185 merged PRs touch
 `engine/hooks/`, 13 of 28 hooks needed post-ship repair, and 37 PRs did
 nothing but repair a shipped detector. `diu-stop` took 8, `pr-schema-gate`
 7, `wrong-check-reflect` 6, `scope-lock` 4. `gh-write-verification` hit two
-separate known kinds in three days (#322, #323, #324). None of those were
+separate known kinds in three days. None of those were
 new problems; each was a kind already fixed in another hook.
 
 ## The playbook