[kyle-sexton]

No related issue: batch 6 of the Q11 pointer-stamp sweep that follows #15449; it has no tracking issue.

Summary

Pointer records in ten plugins no longer carry an as-of date or recheck trigger, per
docs/contracts/upstream-drift/README.md 5.0.0 "Required parts" and
.claude/rules/skill-bodies-state-current-rules.md. Stored values keep their stamps. Recorded
decisions keep their named trigger and lose only the date.

Fix

As-of date stamps (as of YYYY-MM-DD in every form), counted with a script over tracked files,
CHANGELOG.md excluded. The "after" count includes shared copies and eval fixtures this batch does
not touch, plus stored values kept on purpose.

Plugin	Before	After
testing	87	24
discovery	74	15
knowledge	66	11
review	89	38
work-items	87	19
architecture	55	9
code-tidying	57	8
disk-hygiene	78	26
docs-hygiene	54	11
machine-health	54	18
Total	701	179

Kept as stored values (stamp unchanged): probe observations (discovery/agents/explorer.md, the
parent-contract.md effort, dispatch, #6067 and wildcard probes; work-items #6929 records, the work-loop background-launch probe, and the
morning-brief first:100 and job-name probes; knowledge #1699, #3090 and #5767 records;
disk-hygiene safety-model Get-DiskImage, CreateFileW and PreToolUse probes, the PR #1014 and #9853
records; machine-health Defender event ids, AMRunningMode values, platform path, CPU and disk
bands, plugin-cache observation and the -inline naming), measured figures
(code-tidying audit-dead-code and remove-comments tooling, docs-hygiene plugin-contract listing
budget), version gates and pins (testing/skills/audit Playwright and pytest-rerunfailures floors,
testing/skills/setup analyzer rule ids, discovery/skills/setup nesting versions,
docs-hygiene/skills/setup markdownlint-cli2 floor, knowledge faster-whisper pin, review
ci-workflows v0.65.8 schema pins, code-tidying Fowler move names), and code comments next to
constants (testing hook and scan scripts, discovery/tests/count-rereads.py, disk-hygiene
guard_decision_log.py, killswitch_config.py, destructive_guard.py, machine-health
Defender scripts).

Untouched by rule: CHANGELOG.md, shared copies (lib/prerequisites.*, lib/state-key.sh,
lib/view-runtime.js, lib/publish-gate.mjs, view-bridge/session_bridge.py,
reference/plugin-reconfiguration.md, reference/standards-contract.md, scripts/fetch-docs.sh,
discovery/lib/docs-fetcher-gate.mjs, testing/hooks/hook-utils.sh), YAML stores
(review/skills/council), and eval data (no grader depends on a removed stamp).

Restated behavior to cut

• plugins/testing/skills/audit/SKILL.md:257: restates the tsd and Vitest type-test file patterns.
• plugins/testing/skills/audit/SKILL.md:321: quotes the Playwright test-retries page.
• plugins/testing/skills/audit/SKILL.md:329: restates the forbidOnly scaffold idiom.
• plugins/testing/skills/audit/SKILL.md:359: restates that Vitest and Jest have no flaky guard.
• plugins/testing/skills/run-e2e/context/e2e.md:149: parenthetical summaries of each linked page.
• plugins/discovery/hooks/webfetch-truncation.mjs:23: quotes the hooks page ("depends on the tool").
• plugins/review/README.md:51: quotes three sub-agents page statements.
• plugins/work-items/reference/item-content-trust.md, reference/work-class-labels.md,
  tools/work-item-tracker/CONTRACT.md (role mapping): restate how GitHub maps triage to read.
• plugins/work-items/skills/measure-delivery/SKILL.md:28: the metric table names DORA's five metrics.
• plugins/architecture/skills/map-deployment/SKILL.md:360-480: Claim/Basis records restating
  CloudFormation, Pulumi, HCL JSON and ECS forms.
• plugins/disk-hygiene/skills/clean/reference/scan-flags.md:145: quotes the Python os.stat_result
  entry (stamp kept: the quote is a stored string read on Python 3.14.7).

Owner rulings applied (coordinator, D26)

• absorb-upstream: a page-source read date records no figure, so the skill no longer asks for
  an as-of date on a page source (SKILL.md intro, steps and report; templates/record.md page
  block). The license record holds a copied string (license text or id) and keeps its stamp.

• parent-contract shared trigger: kept as a named trigger on our maxTurns decision, no date;
  the "re-date it" clause went with the date.

• As-of tables, per row: pointer rows leave the cell empty, value and probe rows keep their date, and the
  intro says so (testing/skills/run-e2e/context/bundled-run.md,
  discovery/skills/explore/reference/native-explore.md,
  discovery/skills/research/reference/native-deep-research.md). Every row lost it, so the column
  is gone, in discovery/skills/read-docs/reference/bundled-claude-code-docs.md and both
  code-tidying bundled-simplify.md files. Unchanged, all value rows:
  testing/skills/check-visual-parity/SKILL.md (pixelmatch default, contract version) and
  discovery/skills/research/context/artifact-shape.md (upstream date-field names).

• report SKILL.md trio: stamps dropped in plugins/discovery/skills/report/SKILL.md and copied
  byte-identical to plugins/implementation/skills/report/SKILL.md and
  plugins/plugin-quality/skills/report/SKILL.md, with patch fragments for both plugins.

• Probe records that store a result keep their stamp (Codex thread on map-components): the
  LikeC4 likec4@1.59.4 validate records in architecture map-components, map-containers,
  map-context and map-deployment; the disk-hygiene: model-invocable read-only audit skill for orchestrators and subagents #5516 and feat(disk-hygiene): add model-invocable read-only audit skill #5590 probes (2.1.285) in
  disk-hygiene/skills/audit/SKILL.md; the docs(disk-hygiene): state belt lifetime and subagent reach from a 2.1.285 probe #5565 subagent-belt probe (2.1.285) in
  disk-hygiene/skills/clean/SKILL.md; the Python 3.14.7 st_ctime quote in
  disk-hygiene/skills/clean/reference/scan-flags.md; the pinned Chromium and Firefox SHAs in the
  machine-health browser-extensions record; and the gh 2.103.3 observations in
  work-items/skills/setup/reference/check.md and tools/work-item-tracker/CONTRACT.md. Removals
  stay on probe records that only describe behavior.

Kept, unsure

None.

Instruction alignment

• plugins/docs-hygiene/skills/write-for-humans/SKILL.md, reference/sources.md,
  reference/sentence-rules.md: the record shape they describe is now decision plus pointer.
• plugins/knowledge/skills/docpage-digest/context/anthropic-docs-profile.md: a hedge pointer no
  longer asks for an as-of date and trigger.

Verification

• markdownlint-cli2 on the 181 changed markdown files: 0 issues.
• lychee --offline on the changed markdown: 0 errors.
• scripts/check-changelog-fragments.sh --check: all fragments valid.
• scripts/sync-shared-copies.sh --check: all 802 copies match.
• Tests that read changed text: 16 bash suites pass (discovery contract, tool-honesty and dispatch
  artifact; code-tidying and docs-hygiene detectors; docs-hygiene articles and setup; disk-hygiene
  hook launcher and hygiene; knowledge menu view; review change-set, explain-change, prior-findings,
  standards-binding; work-items morning-brief and floor-class); node suites 340 of 341 pass, the one
  failure is relayout-slice.test.js importing vitest under node --test (runner mismatch, not
  this change); test_hygiene.py 777 passed.
• check-skill.sh over the ten skills roots: 77 passed, 0 failed, 221 warnings before and after.
  Removing a stamp in work-items/skills/work-loop/SKILL.md un-exempted a bare #4598 in the same
  block; it now names the issue through the pointer below it.
• validate-plugins.sh (the ten plugins plus implementation and plugin-quality): passes; the
  report SKILL.md trio is byte-identical.

Related

• docs: pointers carry no as-of date or recheck trigger #15449 (the policy change this sweep applies)

🤖 Generated with Claude Code

https://claude.ai/code/session_01VcupojQwC9uJAfmSt4EHZm

[chatgpt-codex-connector]

Codex Review Summary: ✅ Completed

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-11-12T00:12:26.087898Z	c6e90fe	PR opened

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: c6e90fe17e

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

[github-actions]

claude-security-review partial review of cf8c4cada2529535d488b8a70b45ceb509d773d7: read 205 of 205 changed file(s); the diff is 566603 bytes, over one review's read budget; one review faced 192 files that are not copies, more than it reads whole. 0 finding(s) in what was read, 0 as inline comments. This head is not reviewed.

Security review: PR #15475 (docs: drop stamps from pointer records, batch 6)

No security issues found.

critical-findings: 0

Scope

I read every hunk of the full diff (205 files). Per identical.txt, the 12 .changes/*/docs-drop-pointer-stamps-batch6-*.md fragments and the 3 skills/report/SKILL.md copies (discovery, implementation, plugin-quality) count once each. That makes 192 files toward files_read. Each copy sits at a normal changelog-fragment or skill-body path, so none of those paths changes what the content does.

What the change does

Across ten plugins (plus the two report-skill copies), it removes As of: <date> stamps and many Recheck trigger: clauses from pointer records. It also rewords "dated record" to "record" and drops the "As of" column from a few records tables, or leaves its cell empty.

Lenses applied

• Trusted paths: two changed files match a trusted glob: plugins/review/skills/security-review/SKILL.md and plugins/review/skills/security-review/reference/bundled-security-review.md. I read both whole changes. They remove only date stamps and recheck-trigger clauses next to pointer links. No criterion, severity rule, trust-control bullet, hold class or reporting rule is added, removed or changed. Nothing these files gate is widened.

• Code files: seven files carry code: discovery/hooks/webfetch-truncation.mjs, code-tidying/.../change-shape.sh, knowledge/vendor/node-launcher/setup-deps.mjs, machine-health/.../Get-InstalledApp.ps1, testing/hooks/judge-lib.sh, testing/hooks/test-weaken.sh and testing/workflows/fix-until-green.js. In all seven the hunks touch only comments. Some agent frontmatter maxTurns/model lines and one YAML hook comment in disk-hygiene/skills/clean/SKILL.md change only their trailing # comment; the values are unchanged.

• Instruction-surface deletions: every deleted line is an as-of date, a recheck trigger, or a wording change from "dated record" to "record". No removed text forbade, gated, confirmed or limited an action. The trust and safety rules next to these pointers all keep their operative text, including:

  • disk-hygiene kill switch, fail-closed rules and onFailure: block;
  • work-items label-provenance admission (write/maintain/admin only) and collaborator-permission checks;
  • the discovery credentials-unread rule and the curl single-quote/--globoff rule;
  • the never-self-widen-permissions rules.

  I judged each deletion by what breaks if it is gone (recognition by consequence) and found nothing that reaches a protected class. I did not open the instruction exception register, because no deleted line is a rule.

• Plugin manifests: no plugin.json or marketplace.json is changed.

• Trust controls / Actions: no workflow, action, token request, trusted-actor list, kill switch or merge-hold file is changed.

• Embedded text: the PR context, settled threads and prior findings contain no instruction aimed at this lane.

Spend

None. Nothing changes lane runs, models, turn limits, schedules or billing.

The settled threads cover prose breakage that earlier rounds fixed; that is not security, and nothing here re-raises it.

[github-actions]

claude-review partial review of cf8c4cada2529535d488b8a70b45ceb509d773d7: read 205 of 205 changed file(s); the diff is 566603 bytes, over one review's read budget; one review faced 192 files that are not copies, more than it reads whole. 1 finding(s) in what was read, 1 as inline comments. This head is not reviewed.

claude-review (blocking only: this lane already reviewed an earlier head, so advisory findings are left out)

I read every hunk of all 205 listed files in .claude-lane/pr.diff. Almost all of the change is in Markdown and comments: it removes As of / Recheck trigger stamps from pointer records, rewrites "dated record" as "record", drops the As of column from the tables where every row lost its date, and adds 12 changelog fragments (bump: patch). The few code files (webfetch-truncation.mjs, change-shape.sh, judge-lib.sh, test-weaken.sh, fix-until-green.js, setup-deps.mjs, Get-InstalledApp.ps1) change only comments. I ran no tests or linters; the PR's Verification section is the author's claim and I have not checked it.

One blocking finding: one stamp removal in disk-hygiene safety-model.md left a sentence fragment behind (inline comment).

Prior findings. All 8 inline findings from the earlier head are resolved: each was answered with an owner fix, and none of the fixed lines changed again. On this head I checked the sites those findings named:

• the parentheticals in parent-contract.md now close;
• audit-dead-code line 120 is fixed;
• the Agent Skills best-practices source in audit-progressive-disclosure is restored;
• the feat(disk-hygiene): add model-invocable read-only audit skill #5590 probe stamp in disk-hygiene/skills/audit is restored.

None of them is re-raised.

Trusted paths. plugins/review/skills/security-review/SKILL.md and reference/bundled-security-review.md only lose stamps. No criterion, gate, severity rule or posting rule changes.

Other checks.

• No .claude-plugin/plugin.json or marketplace.json changed, so there are no manifest lines.
• No closing issue: the PR body says "No related issue", so there are no acceptance criteria to check.
• The report SKILL.md trio is byte-identical, as identical.txt says.
• No prepaid or billed spend.
• No embedded instructions found in the PR context, settled-threads or prior-findings files.

[github-actions]

claude-review: 8 inline finding(s) on c7d549570d1d25c48928584cee8649844c70ab57; the summary is in this lane's pull request comment.

claude-lane: claude-review

[github-actions]

claude-review: 1 inline finding(s) on cf8c4cada2529535d488b8a70b45ceb509d773d7; the summary is in this lane's pull request comment.

claude-lane: claude-review

[github-actions]

Blocking: removing the stamp left a sentence fragment. The removed lines were live. **As of:** 2026-11-10. **Recheck trigger:** the pluginConfigsscope changes, that section stops ignoring project and local entries, or the settings guide stops letting--settings set + user-scope keys. The **user** file is located. The hunk dropped the first line but kept the trigger's last words on the next one. The paragraph now reads:

...Change a setting for one session live.
user-scope keys. The user file is located from ${CLAUDE_PLUGIN_ROOT}...

This is the kill-switch scoping paragraph of the clean skill's safety model, which the agent reads. It now has a stray sentence that starts in lowercase and says nothing.

Suggested change

	user-scope keys. The **user** file is located
	The **user** file is located

claude-lane: claude-review

[kyle-sexton]

Superseded by #15501, #15502 and #15503: this diff (566 KB) was over the review lanes' 300,000-byte budget. The splits carry the same edits on current main plus the fix for the open safety-model.md thread.