Index of every page on this site: llms.txt
All skills
a2ui-project avatar

/a2ui-audit

@6bf4cc2
by A2UIa2ui-project/a2ui18k stars
1,430
Run it yourself

Main coordination skill to run the blueprint compliance, documentation synchronization, and test quality audits, posting the combined results as a labeled GitHub issue.

Use this Skill: https://skilld.dev/gh/a2ui-project/a2ui/a2ui-audit

Nothing lands on disk. Nothing to clean up.

Fork this Skill

Edit a local copy. It keeps the author and licence.

SKILL.md

≈45 tokens for metadata: the name and description. ≈1.6k when used: this file. ≈2.1k more on demand in 1 file.

Description uses 2.3% of example budget

Before choosing a Skill, your Agent reads its name and description. All available Skills share that space.

  • A shorter description leaves more room for other Skills. This entry exceeds our 1% size suggestion.

In our Claude Code example, all Skill names and descriptions share 8,000 characters. This Skill uses ≈181 characters, or 2.3%.

The 1% threshold is a size suggestion. Longer descriptions can still fit.

Your model, settings, and other Skills decide how much text your Agent can read.

Example settings and source

The example uses a 200k-token context and default Claude Code settings. The count includes the name, description, separators, and when_to_use when present. Codex also counts local file paths.

Skit's source and limits: Codex 0.160.1, Claude Code 2.1.292.

A2UI Compliance Verification Skill

This skill coordinates the execution of specification compliance, documentation sync, and test quality audits, and publishes the combined results as a GitHub issue.

Other than writing a report and creating an issue, this skill is purely a read-only skill; no changes are to be made to the codebase. Do not attempt to fix issues.


Instructions

  1. Audit Codebase Blueprint Compliance:

    • Execute python3 blueprints/skills/a2ui-blueprint-compliance/scripts/check_compliance.py to dynamically discover all 8 production codebases and generate the baseline Markdown status report.
    • Run the sub-skill a2ui-blueprint-compliance to audit each discovered codebase for missing features or blueprint drift.
    • Save the compiled Markdown report to a temporary file in the workspace (e.g., compliance_report.md) under the header ## Codebase Blueprint Compliance Audit.
  2. Audit Code vs. Documentation Synchronization:

    • Run the sub-skill a2ui-doc-sync-check to audit documentation drift across docs/, specification/, codebase READMEs, and inline docstrings.
    • Append the returned Markdown report to compliance_report.md under a new header ## Code & Documentation Sync Audit.
  3. Audit Test Quality & Assertion Strength:

    • Run the sub-skill a2ui-test-quality-check to audit test suites across all codebases for weak assertions and behavioral verification.
    • Append the returned Markdown report to compliance_report.md under a new header ## Test Quality & Assertions Audit.
  4. Summarize the report:

    • Add a ## Summary section at the top of the compliance_report.md file with a detailed overview of the audit scope and key findings across all audited codebases.
    • Add a ## Recommendations section listing actionable, prioritized follow-up items formatted as a numbered list (e.g. 1. **P0**: ...). Each numbered item MUST represent a concrete, self-contained remediation task.
    • Immediately below the recommendations list, include the mandatory automated remediation callout:

      🤖 Automated Remediation: Comment /fix <number> (e.g., /fix 1) on this issue to have an agent create a draft PR for that recommendation.

  5. Format and Detail Requirements:

    • MANDATORY COMPLETE REPOSITORY COVERAGE: The report MUST include status rows and detailed findings for ALL 8 production codebases discovered by check_compliance.py. Do NOT truncate, shortcut, or abbreviate the summary tables or detailed findings.
    • For each audit section, provide specific evidence:
      • Exact file paths (e.g. python/a2ui_agent/transport.py:L45-L60).
      • Specific function/class names, parameter mismatches, or missing feature descriptions.
      • Concrete examples of weak assertions (e.g. assert response is not None in eval/tests/test_strategies.py).
    • Follow the structure provided in the Report Format Template below.
  6. Publish Report:

    • Execute the local Python helper script to create the GitHub issue containing the combined reports:
      python3 .agents/skills/a2ui-audit/scripts/create_compliance_report.py compliance_report.md --repo a2ui-project/a2ui
    • The publishing script will create the GitHub issue, automatically inject copyable remediation prompt blocks referencing the newly created issue URL under each recommendation item, and update the issue body.
    • Ensure the helper script runs successfully.
    • Clean up the temporary file compliance_report.md after completion.

Report Format Template

When compiling compliance_report.md, use the following structure:

## Summary

[Provide a 2-3 paragraph detailed summary explaining the repository state, key areas audited across SDKs, renderers, and tests, and major findings.]

## Recommendations

1. **[Priority]**: **[Title]**
   - [Clear detailed explanation of the fix needed, specifying affected directories or modules.]

2. **[Priority]**: **[Title]**
   - [Clear detailed explanation of the fix needed, specifying affected directories or modules.]

## Codebase Blueprint Compliance Audit

| Codebase Implementation | Associated Module   | Status | Commits Behind | Current Commit | Latest Commit |
| ----------------------- | ------------------- | ------ | -------------- | -------------- | ------------- |
| `path/to/codebase`      | `associated_module` | Status | X              | `commit_hash`  | `latest_hash` |

### Detailed Findings & Discrepancies

- **[Codebase Path]**: [Detailed description of missing specifications, missing required features, or frontmatter commit hash drift.]

## Code & Documentation Sync Audit

| Directory     | Status | Identified Issues |
| ------------- | ------ | ----------------- |
| `path/to/dir` | Status | Brief Summary     |

### Detailed Findings

- **README Mismatches**:
  - `path/to/README.md`: [Exact invalid command or outdated setup step found in README.]
- **Docstring / API Drift**:
  - `path/to/file.ext:L12-L34`: [Specific parameter or return type mismatch between implementation and docstring/comments.]

## Test Quality & Assertions Audit

| Suite / Module  | Status | Observations  |
| --------------- | ------ | ------------- |
| `path/to/tests` | Status | Brief Summary |

### Detailed Findings

- **Weak Assertions**:
  - `path/to/test_file.ext` (`test_function_name`): [Explanation of weak assertion, e.g. using `assertNotNull` instead of schema/type check.]
- **Missing Edge Case Tests**:
  - `path/to/module`: [Boundary conditions, error cases, or invalid inputs that lack test coverage.]

References

Source: SKILL.md on GitHub

skilld matched fixed text patterns in SKILL.md and file names. Patterns miss obfuscated code.

skilld run checks every file with the same patterns. It asks for approval before it loads a Skill with a behavior marked Needs approval.

1 warning2mo3 checks · Risk MEDIUM
  • Gen Agent Trust Hub2mo

    The skill coordinates repository audits and automated remediation tasks using the GitHub CLI and git. It is classified as medium risk due to a contradiction between its stated 'read-only' nature and its actual capability to push branches and create pull requests. Additionally, it is susceptible to indirect prompt injection from the external repositories it audits.

  • Socket2mo

    No alerts

  • Snyk2mo

    Risk: LOW · No issues

Signed by skilld at 6bf4cc2. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub yesterday.

Activeupdated 2 weeks ago
metadata
{
  "internal": true
}

README badge

README badge for a2ui-project/a2ui/a2ui-audit