vc-audit-vc

Validate agent harness consistency across Claude, Codex, README, and protocol files.

Updated Apr 9, 2026
One-click install
npx skills add https://github.com/HyuseCS/project-c --skill vc-audit-vc-hyusecs
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: vc-audit-vc
Source: https://github.com/HyuseCS/project-c/tree/main/.claude/skills/vc-audit-vc
Command: npx skills add https://github.com/HyuseCS/project-c --skill vc-audit-vc-hyusecs

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve? Agent harnesses that span multiple AI tools (Claude, Codex) and documentation surfaces drift out of sync over time: agents lose parity, skills go unregistered, README listings go stale, and protocol references break. This Skill runs a battery of validators to detect and report that drift before it causes broken workflows. ## Core Features & Use Cases - Agent Parity Validation: Compares Claude markdown agents against Codex TOML mirrors for missing files, description mismatches, and mode inconsistencies. - Skill Registry & README Sync: Verifies every skill on disk has valid frontmatter, is discoverable via the .agents/skills symlink, and is listed in README.md. - Protocol & Wiring Audits: Checks that protocol files are referenced correctly, skill invocation tokens resolve to registered skills, and agent frontmatter contains required fields like effort and skills. - Use Case: After renaming an agent or adding a new skill, run this audit to confirm Claude/Codex parity, README listings, and protocol wiring all still line up before committing. ## Quick Start Run the harness audit to check agent parity, skill registry consistency, README sync, and protocol wiring across the repository.

Frequently Asked Questions about vc-audit-vc

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I check Claude and Codex agent parity?

Run the validate-agent-parity.mjs script with Node.js from the repository root. It compares agent names, descriptions, modes, and normalized bodies between .claude/agents markdown files and .codex/agents TOML files, reporting mismatches as JSON.

How to validate that all skills are listed in README.md?

Run validate-guide-sync.mjs, which extracts agent and skill names from README.md sections and cross-checks them against SKILL.md files on disk. Missing disk entries are failures; stale README entries are warnings.

What does the skill registry validator check?

validate-skills.mjs confirms each skill folder has a SKILL.md with name and description frontmatter, verifies the .agents/skills symlink resolves to .claude/skills, and flags naming, length, and stale-wording issues in descriptions.

Why does the audit report unregistered skill references?

validate-skill-invocation-wiring.mjs scans agent bodies for backtick-wrapped vc-* tokens and fails any token not found in generated-skills-catalog.json. Known agent names are allowlisted, so the failure means a referenced skill is missing from the catalog.

Can I run the validators outside a git repository?

Yes. Most validators fall back to the current working directory when git rev-parse fails, so they work in freshly scaffolded projects. Some, like validate-guide-sync.mjs, use process.cwd() directly.

When should I use this audit instead of the context audit?

Use this audit for harness consistency: agent parity, skill registration, README sync, and protocol wiring. For context routing, grouping, and discoverability audits, the SKILL.md explicitly delegates to the separate audit-context skill.