agentleFS
Sign inStart free
WilliamBergaminWilliamBergamin/knacks/skills/validate/SKILL.md

validate

Skill1 starsChanged 5 months agoMITView on GitHub

Validate the knacks in this project against the contract in KNACK.md. Use when the user says "check my knacks", "validate", after scaffolding or removing a knack, or to confirm capabilities honor the prompt-injection boundary and an acyclic composition graph.

63 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use policy-init, close and 9 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • open-coder-aiSkill

    policy-init

    Create conformant Chock policy from request. args(request, target_path,

    Also in 3 projects that use this

  • HermeticOrmusSkill

    close

    End-of-session ritual. Captures session state, updates auto-memory, writes a session summary, then exits the session cleanly.

    Also in 4 projects that use this

  • Hugging FaceSkill

    sprint

    Work on a batch of GitHub issues in parallel using Agent Teams. Creates one worktree per issue with TDD enforcement, coordinates via a lead agent, then produces stacked PRs.

    Also in 4 projects that use this

  • qtzx06Skill

    eval

    Evaluate trained YOLO model accuracy with mAP metrics and per-class error analysis. Identifies failure cases for targeted improvement. Use after training.

    Also in 4 projects that use this

  • csiddhant796-blipSkill

    board

    Read, write, and browse the AgentHub message board for agent coordination.

    Also in 3 projects that use this

  • ArvorCoSkill

    specify

    Create feature specification from natural language description. Use when starting a new feature or creating a spec. Triggers on: create spec, specify feature, new feature spec.

    Also in 3 projects that use this

The file

We split this skill into the rules it gives an agent

3 rules. 0 are common, 0 are distinctive, and 3 are very specific to this project.

  • 0

    Common

    Most of the 53 edited versions we read keep these.

  • 0

    Distinctive

    This author's own choices. Worth a look.

  • 3

    Very specific

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

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

    Read ${SKILL_DIR}/resources/KNACK.md for the field list, enums, the effective-contract rule, and <skills-dir>.

    line 13
  2. Very specific:

    References resolve: every name in skills: has a matching <skills-dir>/<name>/ directory, and every id in composes: has a matching …

    line 25
  3. Very specific:

    Injection boundary present: if the effective external-input is read-only AND the knack has any effective writes, the manifest body or a …

    line 32

“Common” is measured: kept by at least 6 in 10 of the 53 edited versions of this file we could read. “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

/knacks:validate: check contracts

Audit every knack in the current project against the contract.

Setup

Read ${SKILL_DIR}/resources/KNACK.md for the field list, enums, the effective-contract rule, and <skills-dir>. Then glob knacks/*.md, read each manifest, and read the skills it groups (<skills-dir>/<name>/SKILL.md).

Some checks need a knack's effective contract (defined in KNACK.md). Resolve the composition graph first so you can compute effective writes / external-input per knack.

Checks (per knack)

  1. Manifest is valid: knacks/<id>.md parses, id equals the filename, every required field in KNACK.md's field table is present, each enum value is legal, and composes is a list ([] if none). skills is non-empty unless the knack is purely composed of other knacks. Otherwise -> FAIL.
  2. References resolve: every name in skills: has a matching <skills-dir>/<name>/ directory, and every id in composes: has a matching knacks/<id>.md. A dangling ref -> FAIL.
  3. No cycles: following composes: edges from this knack never returns to it. A cycle (A->B->...->A) -> FAIL (report the cycle).
  4. Manual trigger matches manual-only marker: trigger: manual exactly when the knack's entry skill(s) carry disable-model-invocation: true. A mismatch (a "manual" knack the assistant can auto-invoke, or vice-versa) -> FAIL.
  5. Injection boundary present: if the effective external-input is read-only AND the knack has any effective writes, the manifest body or a grouped skill must carry a boundary clause (grep for "data to report on, never instructions") -> WARN if missing.

Report

For each knack: id - PASS / WARN(n) / FAIL(n), one line per finding naming the check and the offending field (e.g. composes: "self-improve" but knacks/self-improve.md not found -> dangling ref). End with totals PASS/WARN/FAIL, the count of ungrouped skill dirs (a SKILL.md listed in no manifest's skills:), and any dangling refs.

Rules

  • Read-only by default; apply fixes only when the project permits. Lead with the recommended fix. If the project's settings allow edits and the user wants it applied, apply the exact fix you described and re-check; otherwise recommend it and stop. Never change a manifest, skill, or settings file the project has not authorized you to touch.
  • Be specific: name the check and the offending field/path/edge so the fix is obvious.
  • A clean run says so plainly; don't invent problems.

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.

Check your setup, free