agentleFS
shinprshinpr/codex-workflows/.agents/skills/recipe-quality-profile/SKILL.md

recipe-quality-profile

Skill39 starsChanged 12 days agoMITView on GitHub

Proposes and, after user confirmation, creates or maintains docs/project-context/quality.yaml from repository-specific review requirements. Use when asked to generate or update a repository quality profile.

1 other project uses a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Before loading required skills, use wc -c to size the files and set both exec_command.max_output_tokens and the outer // @exec: …

    line 8
  2. Distinctive:

    When the user's response changes a proposed criterion or resolves a policy choice, repeat its repository comparison and present the revised …

    line 45
  3. Very specific:

    Read the resulting YAML and verify version 1, unique IDs, all four dimension fields, positive observable pass conditions, readable evidence …

    line 46
  4. Distinctive:

    When no repository-specific dimension survives this process and no profile exists, report that no profile content is supported and leave …

    line 48

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Required Skills [LOAD BEFORE EXECUTION]

Before loading required skills, use wc -c to size the files and set both exec_command.max_output_tokens and the outer // @exec: {"max_output_tokens": N} budget to accommodate their contents, including the combined output of batched reads. Split oversized reads upfront; recover any truncated content before proceeding.

  1. llm-friendly-context — make each generated condition unambiguous, executable, and observable
  2. coding-rules — distinguish repository-specific policy from general code quality knowledge

Purpose

Establish repository-specific code-quality acceptance conditions with the user, then create or update docs/project-context/quality.yaml. Create docs/project-context/ when the directory is absent.

Target repository and, when a profile already exists, requested policy change: $ARGUMENTS

Profile Contract

Use this shape:

version: 1
review_dimensions:
  - id: stable-kebab-case-id
    applies_when: Observable condition that makes this repository rule relevant to a change.
    pass: Observable accepted state to verify.
    evidence:
      - "repository/path: section, identifier, or contract"

Each dimension represents one repository-specific decision. id identifies it across reviews, applies_when limits its review surface, pass defines acceptance, and evidence records why the repository owns the rule.

Authoring Flow

  1. Resolve the target repository and read its existing docs/project-context/quality.yaml when present. When a profile exists without a requested policy change, ask for the intended change and keep the profile unchanged.
  2. Establish candidate dimensions before deciding profile content:
    • For initial creation, derive candidates from acceptance conditions that repository instructions, contributor documentation, CI, manifests and scripts, schemas and public contracts, tests, or established implementation patterns express or enforce.
    • For an update, translate the requested policy change into candidate additions, changes, or removals. Match it to the current dimension that owns the same policy when one exists and preserve other dimensions as the proposed unchanged set.
  3. State a candidate applies_when and pass, then identify the repository claims that must hold for that criterion to mean what it says. Inspect supporting and contradicting evidence, following declarations, workflows, contracts, tests, and usages wherever their difference could change the candidate's applicability, accepted state, or ownership.
  4. Compare the candidate with the current profile and repository evidence. Record duplicated policy ownership, contradictory accepted states under overlapping applies_when conditions, conflicting evidence, and current accepted behavior that the candidate would newly reject. Separate repository facts from policy choices that only the user can make.
  5. Retain a candidate for the proposal only when its failure would change implementation acceptance and every repository fact it depends on is supported by cited evidence. Present remaining policy choices for user confirmation. When a required repository fact cannot be established, report the exact evidence needed and omit that candidate from the proposal. A general language or framework concern qualifies when repository evidence adds a distinct local rule. Express each retained dimension as one positive observable pass condition with the narrowest useful applies_when boundary, and consolidate dimensions that would produce the same finding and correction.
  6. Before writing, present the proposed additions, changes, and removals; confirm that other dimensions remain unchanged; show the supporting and contradicting evidence for each proposed modification; and state every unresolved policy choice with its effect on review acceptance. [STOP — BLOCKING] Keep the repository unchanged until the user explicitly confirms a proposal with no unresolved choices.
  7. When the user's response changes a proposed criterion or resolves a policy choice, repeat its repository comparison and present the revised proposal for confirmation. After confirmation, write only the confirmed profile content.
  8. Read the resulting YAML and verify version 1, unique IDs, all four dimension fields, positive observable pass conditions, readable evidence references, and consistency with the confirmed proposal.

When no repository-specific dimension survives this process and no profile exists, report that no profile content is supported and leave the repository unchanged.

Result

Before confirmation, report:

  • proposed additions, changes, and removals, plus confirmation that other dimensions remain unchanged;
  • supporting and contradicting repository evidence for each proposed modification, plus exact evidence needed for any candidate that cannot yet be proposed;
  • policy choices requiring the user's decision and how each choice changes review acceptance.

After confirmation, report:

  • profile path and whether it was created, updated, or left unchanged;
  • dimensions added, changed, or removed;
  • evidence used for each changed dimension;
  • an exact validation limitation when the profile could not be verified.

Completion Check

  • llm-friendly-context was read before authoring criteria
  • Every dimension changes a review decision and cites repository or user-policy evidence
  • Conditions are positive, observable, and limited by applies_when
  • The profile contains only repository-specific acceptance conditions
  • Candidate criteria were compared with supporting and contradicting repository evidence
  • The user confirmed the proposed profile content before any repository write
  • Dimensions outside the requested update remain unchanged and the resulting set is internally consistent
  • The written profile satisfies the Profile Contract

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.