fix(authoritative-text-rules): single-property Scope membership - #201
Merged
Merged
Conversation
…d-as-instructions property alone The second Scope property (describes no separate artifact) excluded script-wrapping skill bodies from every item despite their being agent-executed instructions. Membership now turns on the first property alone; claims describing a co-located artifact route to claim-vs-referent items on their own triggers. Refs #153
…cope Review fixes on the new Scope text: state that applying a definition file by reference is execution, mark the wrapped-script examples as illustration, route referent checks to quality-list by name, attach the separate-artifact clause to the claim rather than the file, and state that an item whose sweep reaches such a claim still reports on its own concern conditions. Refs #153
There was a problem hiding this comment.
Pull request overview
Clarifies the authoritative-text-rules Scope predicate to use a single membership property (“executed as instructions”), explicitly including script-wrapping skill bodies, and routes claim-vs-referent checks to quality-list based on its own triggers. Also bumps the plugin version to reflect the behavior-definition change.
Changes:
- Redefines authoritative-text membership to depend solely on “agent executes it as instructions,” including definition-by-reference.
- Clarifies that referent-checking for claims about separate artifacts belongs to
quality-list, with a noted carve-out where anauthoritative-text-rulesitem’s sweep reaches such claims. - Bumps
.claude-plugin/marketplace.jsonversion from2026.8.6to2026.8.7.
Reviewed changes
Copilot reviewed 2 out of 2 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
| skills/authoritative-text-rules/SKILL.md | Updates Scope definition and boundary text to use a single executed-as-instructions membership property and clarify routing to quality-list. |
| .claude-plugin/marketplace.json | Version bump to 2026.8.7. |
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
…ile for which it holds'
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
The Scope section of
authoritative-text-rulesdefined authoritative text by two properties, and the second — "it describes no artifact separate from itself" — admitted a strict and a loose reading that disagree on script-wrapping skill bodies, leaving their membership undecided. Membership now turns on the first property alone: text an agent executes as instructions qualifies in full, wrapped-script describers included, and checking a claim about a separate artifact against its referent belongs toquality-list, on its items' own triggers. Closes #153Changes
skills/authoritative-text-rules/SKILL.md, Scope section: membership is decided by the single executed-as-instructions property, with applying a definition file by reference counting as execution. Claims about a separate artifact route toquality-list, with an explicit carve-out for anauthoritative-text-rulesitem whose own sweep reaches such a claim —case-space-totality's does. The source-code exclusion rests on "a machine executes it", the docstring exclusion on "read rather than executed", and the sentence listing the recurring instances now names script-wrapping skill bodies..claude-plugin/marketplace.jsoncarries version 2026.8.7 (bumped once on the branch, per the squash-merge rule).Impact
done-check,todo-check,research,review-pipeline,finding-to-audit) apply the Scope section by reference and need no edits; their existing one-property paraphrases —done-check's "text an agent executes as instructions",finding-to-audit's "text the agent executes as instructions" — now match the predicate exactly.done-checkdispatches now unambiguously includes the script-wrapping skill bodies (stdlib-audit,gh-body-check,copilot-review) in scope.research's rule-set baseline can no longer skip its read of the listed item bodies for changes touching those bodies, since that skip requires every touched surface to be excluded by the Scope section.Verification
done-check's auditor shape, applying every item inauthoritative-text-rules' Items index to this diff, returned one row per indexed item and no concerns.rgover the repository for verbatim substrings of the removed lines ("both of these properties", "describes no artifact", "no referent its claims") returns no hits outside the rewritten file.mdformatpre-commit hook passed on both commits.Notes
quality-list, whose current triggers do not reach skill-body prose; Skill-body claims about co-located artifacts escape diff-time gates #200 tracks the trigger widening.