audit

Validate skills against mechanical checks: structure (S1–S5), quality (Q1–Q5, Q10–Q11), and security (E1–E2, E6, E9).

Commands

audit validate

# Validate all skills in the project
npx cyberplace@<version> audit validate

# Validate a single skill
npx cyberplace@<version> audit validate --path skills/my-skill

# Output for agents
npx cyberplace@<version> audit validate --format agent

# Output for scripts
npx cyberplace@<version> audit validate --format json

Options:

Flag	Description
--path <path>	Validate a single skill directory or SKILL.md file
--root <path>	Repo root (defaults to current directory)
--format <format>	text (default), agent, or json

What it checks

Range	Category	Severity
S1–S5	Structure	CRITICAL–MEDIUM
Q1–Q5, Q10–Q11	Quality	HIGH–MEDIUM
E1–E2, E6, E9	Security	CRITICAL–HIGH

Full review (Q6–Q13, E3–E8, P1–P3) requires the improve-skill agent skill (aced plugin).

CI integration

- run: npx cyberplace@$(npm view cyberplace version) audit validate

The command exits non-zero on any CRITICAL or HIGH finding, making it safe to use as a blocking CI check.