Skip to content

docs: cut restated upstream behavior left by the stamp sweep (batch 6) #15514

Open

Description

@kyle-sexton

Problem

The stamp sweep in #15475 (docs: drop stamps from pointer records (batch 6)) found these passages restating upstream behavior. Under contract 6.0.0 (#15468) restated behavior is not stamped; it is cut to a pointer or rewritten as this repository's own rule. The sweep left them unchanged because cutting them changes what a skill tells the agent.

Items

  • 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.

Do

Per item: cut to the anchored pointer, keep as our own rule with its reason, or record why it stays. A cut that changes agent behavior needs an eval or a stated basis.

Related

Refs: #15475, #15468

Activity

  1. added
    needs-triageNot yet classified. Floor until a type and one priority tier are set.
    on Nov 12, 2026
  2. github-actions commented on Nov 12, 2026

    @github-actions
    Contributor

    Size: Medium
    Complexity: Low
    Uncertainty: Known
    Flags: none
    Blocked by: none

    Understood: this is a docs cleanup. The stamp sweep in #15475 left a list of passages across several plugins that restate upstream behavior. Each should be cut to an anchored pointer, rewritten as this repository's own rule, or recorded as intentionally kept. The issue says a cut that changes agent behavior needs an eval or a stated basis.

    Nothing is missing. The sibling issues for other batches (e.g. batch 5 and batch 7) cover different files, so this isn't a duplicate of them. I found no merged pull request that already does this work.

    Automated triage. A maintainer reviews it.

  3. removed
    needs-triageNot yet classified. Floor until a type and one priority tier are set.
    on Nov 12, 2026
  4. melodic-automation-lanes commented on Nov 12, 2026

    @melodic-automation-lanes
    Contributor

    Not admitted to the issue lane: no trusted-human acceptance criteria: a listed human must put the confirmed - [ ] criteria in the body or post them as their own comment, or, on an issue a listed human filed, intake's newest refinement comment must hold them unedited. Admission takes the issue once each of those is met, when intake triages it again (.github/actions/resolve-config/admission-rules.json).

  5. melodic-automation-lanes commented on Nov 12, 2026

    @melodic-automation-lanes
    Contributor

    Refinement. A read-only run explored this repository, searched its issues and pull requests, and read the allowed docs to draft what follows. On an issue a listed human filed, it counts as the issue's spec once posted, while unedited; it is still model text, so check it.

    Intended outcome: Each listed passage is cut to an anchored pointer, rewritten as this repository's own rule, or recorded as kept, with a stated basis for any cut that changes agent behavior.

    Related work

    Open questions

    1. For each of the 11 item groups, should the run cut to a pointer, rewrite as our own rule, or keep? The issue says choose per item and decides none.
      RECOMMENDED: Cut to an anchored pointer by default. Keep as our own rule only where agent behavior depends on it, such as the forbidOnly check in the audit skill. Basis: plugins/testing/skills/audit/SKILL.md:47
    2. What counts as the stated basis or eval for a cut that changes agent behavior? An unattended run cannot run evals.
      RECOMMENDED: Accept a one-line basis in the PR body naming the upstream page that still says the same thing, with no eval. Basis: judgment
    3. Should the body be rewritten as a task list with one checkbox per item group so each can be checked?
      RECOMMENDED: Yes, one checkbox per item group, as a single docs-only change. Basis: judgment
    4. Cited line numbers have drifted: the tsd text is at audit/SKILL.md:259 and forbidOnly at :331, and no 'depends on the tool' text was found in webfetch-truncation.mjs. Should the run locate each passage by its content and skip items already gone?
      RECOMMENDED: Yes, locate by content, and record in the PR any item whose passage no longer exists. Basis: plugins/testing/skills/audit/SKILL.md:259

    Sources read

    • plugins/testing/skills/audit/SKILL.md:47: rule-only-not-forbidden check relies on forbidOnly
    • plugins/testing/skills/audit/SKILL.md:259: tsd README pointer with as-of stamp
    • plugins/testing/skills/audit/SKILL.md:331: forbidOnly: !!process.env.CI scaffold idiom
    • plugins/discovery/hooks/webfetch-truncation.mjs: no 'depends on the tool' match found

    To answer, post a comment that starts with this line:

    **Decision recorded.** <answer>
    

    Automated refinement by the lanes App. Answer with a comment whose first line is **Decision recorded.**.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    Fields

    Work status

    needs-owner-decision

    Priority

    low

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions