agentleFS
QuirkybrainQuirkybrain/claude-code-swe-framework/.claude/skills/artifact-validation/SKILL.md

artifact-validation

Skill1 starsChanged 53 days agoMITView on GitHub

Validate an existing or newly produced Artifact for completeness, consistency, correctness, freshness, dependency compatibility, and target fitness. Use before reuse, handoff, downstream consumption, or a Gate decision.

The file

We split this skill into the rules it gives an agent

11 rules. About 0 are common, 9 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    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. Distinctive:

    Run before consuming an existing Artifact, after an Artifact is created or changed, when an upstream dependency changes, or when prior …

    line 14
  2. Distinctive:

    Verify content-based Artifact classification and provenance.

    line 30
  3. Distinctive:

    Check completeness against its intended consumer and Definition of Done.

    line 31
  4. Distinctive:

    Check internal consistency and consistency with authoritative Artifacts.

    line 32
  5. Distinctive:

    Check correctness using source evidence, executable checks, or independent review appropriate to type.

    line 33
  6. Very specific:

    Check freshness against upstream revisions, repository state, and Last Verified.

    line 34
  7. Distinctive:

    Check dependency compatibility, including versioned Contracts, schemas, constraints, and assumptions.

    line 35
  8. Distinctive:

    Check fitness for the current target.

    line 36
  9. Very specific:

    Never infer VALID from existence, format quality, author confidence, or an old PASS.

    line 47
  10. Distinctive:

    Use the Clarification Gate for authority or intent conflicts.

    line 67
  11. Distinctive:

    Stop after recording a supported disposition and state, when evidence is insufficient and explicitly blocked, or when user authority is …

    line 71

≈ “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

Artifact Validation

Purpose

Decide whether a specific Artifact revision can safely support the declared target and downstream use, applying Reuse Before Regenerate.

Trigger

Run before consuming an existing Artifact, after an Artifact is created or changed, when an upstream dependency changes, or when prior evidence is stale or incomplete.

Inputs

  • Artifact ID, type, content, revision, source, owner, and intended use.
  • Required upstream Artifact revisions and applicable Project Constraints.
  • Target Stage/Artifact, effective Workflow Level, prior validation, and known limitations.

Prerequisites

  • Confirm the exact revision and validation scope.
  • Resolve or explicitly record uncertain Artifact type and provenance.
  • Identify the minimum validation depth required by target risk.

Procedure

  1. Verify content-based Artifact classification and provenance.
  2. Check completeness against its intended consumer and Definition of Done.
  3. Check internal consistency and consistency with authoritative Artifacts.
  4. Check correctness using source evidence, executable checks, or independent review appropriate to type.
  5. Check freshness against upstream revisions, repository state, and Last Verified.
  6. Check dependency compatibility, including versioned Contracts, schemas, constraints, and assumptions.
  7. Check fitness for the current target; an Artifact may be valid for one scope but insufficient for another.
  8. Record evidence, validator, evaluated revisions, coverage, limitations, and resulting status.
  9. Select exactly one disposition: REUSE, SUPPLEMENT, REVALIDATE, or REGENERATE.

Decision Rules

  • REUSE: evidence is current and the Artifact is complete, consistent, correct, compatible, and sufficient for this target.
  • SUPPLEMENT: useful content remains and bounded gaps can be repaired without replacing valid portions.
  • REVALIDATE: content may still be correct, but freshness or evidence is insufficient.
  • REGENERATE: evidence proves invalidity, or repair is riskier/costlier than replacement.
  • Map the result to VALID, PARTIAL, UNVERIFIED, STALE, or INVALID; do not turn every failed check into INVALID.
  • Never infer VALID from existence, format quality, author confidence, or an old PASS.

Outputs

  • Validation Record with scope, checks, evidence, dependency revisions, result, and limitations.
  • One disposition: REUSE, SUPPLEMENT, REVALIDATE, or REGENERATE.
  • Updated Artifact status and Last Verified when justified.
  • Downstream impact and required follow-up.

Failure Cases

  • Required source or dependency revision is missing.
  • Correctness cannot be checked with available evidence.
  • The Artifact mixes incompatible revisions or authorities.
  • Validation tooling or environment is unavailable.

Return BLOCKED or UNVERIFIED with missing evidence; never fabricate PASS.

Escalation

Use the Clarification Gate for authority or intent conflicts. Use Change Impact when the Artifact or its basis changed. Escalate specialized correctness checks to the responsible Agent while preserving this validation scope.

Stop Conditions

Stop after recording a supported disposition and state, when evidence is insufficient and explicitly blocked, or when user authority is required. Do not regenerate as part of this Skill unless separately authorized.

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.