# Kit Verify

> (ePost) Use when preparing a release, running pre-init checks, auditing kit health, or when asked to "verify", "audit kit", "pre-release check"

- Skill: `majiayu000/kit-verify` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add majiayu000/kit-verify`
- Raw SKILL.md: https://api.skillmd.com/api/skills/majiayu000/kit-verify/raw
- Safety review: pending (external: skill-scanner PASS, skillspector PASS)
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Security
- Author: majiayu000 (https://skillmd.com/u/majiayu000)
- Updated: 2026-09-09
- Page: https://skillmd.com/skills/majiayu000/kit-verify

---


# Kit Verification Workflow

Pre-release/pre-init audit pipeline. Ensures kit integrity before `epost-kit init` or release.

## When to Use

- Before running `epost-kit init` on a project
- Before creating a release/tag
- After batch-editing skills, agents, or connections
- When CI gate needs to validate kit health

## CLI Command

```bash
epost-kit verify            # full audit, errors block
epost-kit verify --strict   # warnings also block (CI mode)
```

## What It Checks

1. **Reference validation** — all agent/skill/command refs point to valid targets
2. **Connection integrity** — extends/requires/enhances targets exist, no cycles, bidirectional conflicts
3. **Frontmatter completeness** — skills have description, keywords, platforms
4. **Package.yaml sync** — skills on disk match provides.skills declarations
5. **Skill-index staleness** — index count matches actual SKILL.md count
6. **Orphan detection** — skills not referenced by any agent or connection
7. **Dependency graph** — auto-generates `docs/skill-dependency-graph.md` (mermaid)

## Exit Codes

- `0` — pass (may have warnings/info)
- `1` — fail (errors found, or `--strict` with warnings)

## Integration

- Wire as pre-commit hook: `epost-kit verify --strict`
- Wire in CI: `npx epost-kit verify --strict --dir .`
- `epost-kit init` can optionally run verify before generating `.claude/`