docs: replace restated upstream facts with pointers in code-tidying, planning, repo-fleet-hygiene #12692

Refs: #7292

Outcome

28 passages in 11 files state a fact an external source owns. Rewrite each as a pointer to the live upstream section, or remove it, so the next upstream change cannot leave a stale copy here. Plugins: code-tidying, planning, repo-fleet-hygiene.

Done when

• Every row below is fixed: the restated value is deleted, the local decision stays in this repository's own words, and the pointer goes to the exact upstream section with an as-of date and a recheck trigger, per .claude/rules/skill-bodies-state-current-rules.md. A record beside a restated value does not fix it. A row marked Remove has no local dependency on the fact.
• The PR passes check-stamps.sh --trigger-less (plugins/attribution/skills/audit/scripts/check-stamps.sh) on the files it touches.
• Each touched plugin has a changelog fragment.

check-stamps.sh also flags a trigger-less stamp at plugins/planning/skills/design/SKILL.md:393 on the same commit; write its recheck trigger if the PR touches that file.

Rows

code-tidying

• plugins/code-tidying/skills/audit-dead-code/SKILL.md:122-126 (r047). Pointer: https://code.claude.com/docs/en/tools-reference#lsp-tool-behavior Drop the includeDeclaration and import-counting behavior claims; keep LSP findReferences as optional extra evidence for one candidate, and point at the LSP tool section. Same text also at: plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:31; a row of its own covers it only if listed.
• plugins/code-tidying/skills/audit-dead-code/SKILL.md:141-142 (r048). Pointer: https://github.com/jendrikseipp/vulture#exit-codes Drop the per-tool exit-code values; keep the rule that exit codes never prove run health, and point each detector at its docs (vulture exit codes, knip CLI, gopls).
• plugins/code-tidying/skills/audit-dead-code/SKILL.md:208-211 (r049). Pointer: https://staticcheck.dev/docs/configuration/#ignoring-problems-with-linter-directives Drop the measured gopls and //lint:ignore claims; keep the decision that Go and grep verdicts live in the report and a declaration comment, and point at staticcheck directives. Same text also at: plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:92, plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:99; a row of its own covers it only if listed.
• plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:31-33 (r050). Pointer: https://code.claude.com/docs/en/tools-reference#lsp-tool-behavior Drop the includeDeclaration and import-counting claims; keep that LSP is one more source, never a substitute, and point at the LSP tool section. Same text also at: plugins/code-tidying/skills/audit-dead-code/SKILL.md:124; a row of its own covers it only if listed.
• plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:69-79 (r051). Pointer: https://knip.dev/reference/configuration#ignore Drop the knip.config.ts jiti note and config-location claims; keep emitting the paste-ready ignore snippet, and point at the knip ignore section.
• plugins/code-tidying/skills/audit-dead-code/context/adjudication.md:81-86 (r052). Pointer: https://github.com/jendrikseipp/vulture#whitelists Drop the vulture whitelist semantics; keep emitting the paste-ready whitelist entry with the local adjudication comment, and point at the vulture whitelists section.
• plugins/code-tidying/skills/batch-simplify/SKILL.md:62-67 (r053). Pointer: https://code.claude.com/docs/en/skills#how-injected-commands-run Drop the one-shell-invocation and worktree-refusal claims and the record link; keep the decision to run these as separate body Bash calls, and point at the section.
• plugins/code-tidying/skills/batch-simplify/SKILL.md:90-94 (r054). Pointer: https://git-scm.com/docs/git-log#commit-limiting (confirm the section) Drop the claim about what git's --since parser requires; keep the conversion table that rewrites h, d, w into long units, and point at git-log's --since option.
• plugins/code-tidying/skills/tidy/lanes/self-update.md:3 (r056). Remove. Delete the sentence about how consumers receive updates; this lane is maintainer-only and nothing local depends on the consumer update path.
• plugins/code-tidying/skills/tidy/lanes/self-update.md:51 (r057). Pointer: https://code.claude.com/docs/en/skills#frontmatter-reference Drop the claim about what a broken fence does to skill discovery; keep the decision to run the fence check from the plugin root, and point at the frontmatter section.
• plugins/code-tidying/skills/tidy/reference/exclusions.md:21 (r058). Pointer: https://code.claude.com/docs/en/hooks#hook-locations Drop the list of plugin hooks.json and frontmatter wiring sites; keep that a script wired as a hook anywhere is excluded, and point at hook locations.
• plugins/code-tidying/skills/tidy/reference/exclusions.md:125 (r059). Pointer: https://code.claude.com/docs/en/plugins/manifest-reference#user-configuration Drop the enumerated userConfig types; keep that the option is a scalar and a per-path list belongs in the repository file, and point at the section.

planning

• plugins/planning/skills/design/SKILL.md:21-25 (r373). Pointer: https://code.claude.com/docs/en/skills#how-injected-commands-run Drop the worktree-refusal and one-invocation claims; keep the decision to use body Bash calls and point at the section. Same text also at: plugins/planning/skills/plan/SKILL.md:22, plugins/planning/skills/prd/SKILL.md:21, plugins/planning/skills/interview/SKILL.md:26-31; a row of its own covers it only if listed.
• plugins/planning/skills/design/SKILL.md:341 (r374). Pointer: https://mermaid.js.org/syntax/c4.html#c4-diagrams Drop the 'experimental' characterization of Mermaid C4; keep the decision to use LikeC4 or C4-PlantUML for the system scope and point at the Mermaid C4 page.
• plugins/planning/skills/draft-goal-condition/SKILL.md:127-130 (r368). Pointer: https://code.claude.com/docs/en/goal#how-evaluation-works Drop the description of what the evaluator can and cannot read; keep the rule that every condition claim must be demonstrable in Claude's output and point at the section.
• plugins/planning/skills/draft-goal-condition/SKILL.md:184-185 (r369). Pointer: https://code.claude.com/docs/en/goal#how-evaluation-works Drop the evaluator's model, context and tool description; keep the instruction that move 3 needs verification performed visibly in the transcript and point at the section.
• plugins/planning/skills/draft-goal-condition/SKILL.md:290-291 (r371). Pointer: https://code.claude.com/docs/en/goal#write-an-effective-condition (confirm the section) Drop the statement about what the app documents on over-limit submission; keep the decision to rely on the pre-submission counter and not assume trimming, and point at the section.
• plugins/planning/skills/draft-goal-condition/SKILL.md:297-299 (r372). Pointer: https://code.claude.com/docs/en/goal#set-a-goal (confirm the section) Drop the permissions-behavior statement; keep the instruction to flag a goal whose check is a shell command and point at the section.
• plugins/planning/skills/interview/SKILL.md:26-31 (r365). Pointer: https://code.claude.com/docs/en/skills#how-injected-commands-run Drop the harness-behavior and worktree-refusal claims; keep the decision to use separate body Bash calls and point at the section. Same text also at: plugins/planning/skills/plan/SKILL.md:22, plugins/planning/skills/prd/SKILL.md:21, plugins/planning/skills/design/SKILL.md:21-25; a row of its own covers it only if listed.
• plugins/planning/skills/plan/SKILL.md:22 (r353). Pointer: https://code.claude.com/docs/en/skills#how-injected-commands-run Drop the claim about how the pre-compute block runs and the worktree refusal; keep the decision to use individual body Bash calls and point at the section. Same text also at: plugins/planning/skills/prd/SKILL.md:21, plugins/planning/skills/interview/SKILL.md:26-31, plugins/planning/skills/design/SKILL.md:21-25; a row of its own covers it only if listed.
• plugins/planning/skills/plan/SKILL.md:30 (r354). Pointer: https://code.claude.com/docs/en/permission-modes#analyze-before-you-edit-with-plan-mode Drop the description of built-in plan mode as a read-only permission mode; keep the statement that this skill is a separate planning discipline and point at the section.
• plugins/planning/skills/plan/SKILL.md:317,320 (r357). Pointer: https://code.claude.com/docs/en/commands#all-commands Drop the /plan argument forms and the who-may-invoke-it behavior; keep the routing decision to offer /plan alongside this skill and point at the commands list.
• plugins/planning/skills/plan/SKILL.md:330 (r358). Pointer: https://code.claude.com/docs/en/sub-agents#built-in-subagents Drop the description of the Plan subagent's behavior and tool limits; keep the routing decision between it and this skill and point at the section.
• plugins/planning/skills/prd/SKILL.md:21 (r360). Pointer: https://code.claude.com/docs/en/skills#how-injected-commands-run Drop the claim about how the pre-compute block runs and the worktree refusal; keep the decision to use body Bash calls and point at the section. Same text also at: plugins/planning/skills/plan/SKILL.md:22, plugins/planning/skills/interview/SKILL.md:26-31, plugins/planning/skills/design/SKILL.md:21-25; a row of its own covers it only if listed.

repo-fleet-hygiene

• plugins/repo-fleet-hygiene/skills/audit/SKILL.md:124-129 (r404). Pointer: https://git-scm.com/docs/git-worktree#list-output-format Drop the claim about list ordering and the show-toplevel behavior; keep the decision to identify the canonical checkout as the main worktree via worktree list porcelain, with a pointer.
• plugins/repo-fleet-hygiene/skills/audit/SKILL.md:147-148 (r403). Pointer: https://docs.github.com/en/graphql/overview/rate-limits-and-query-limits-for-the-graphql-api#primary-rate-limit Drop the node ceiling and hourly point figures; keep the measured cost of 1 per call and point at the primary rate limit and node limit sections.
• plugins/repo-fleet-hygiene/skills/audit/SKILL.md:336-340 (r406). Pointer: https://github.com/git-for-windows/git/wiki/Symbolic-Links (confirm the section) Drop the claim that Windows junctions test as symlinks under Git Bash; keep the discovery-symlink-skip UNKNOWN disclosure decision and point at the page.
• plugins/repo-fleet-hygiene/skills/audit/SKILL.md:377 (r407). Pointer: https://git-scm.com/docs/git-config (confirm the section) Drop the description of what core.bare does and the named remedy command; keep the manual-review disposition and the finding name, and point at the core.bare entry.

Notes

Lines are as of 3bd42cabfe (https://github.com/melodic-software/claude-code-plugins/blob/3bd42cabfe6bdef37a05778d396869af0381bb51); locate each passage by its text if main has moved. Each rNNN is that row's number in the order of #7292's "Findings by plugin". Every row was upheld by three blind judges and survived a refutation pass, run report-only for #12072. Pointers are candidates: fetch each before writing it, and confirm the section still documents the fact.

Parent: #7292