v1.2.0 — Quick audit, auto-trigger skill, hardened prompts

New

• /audit-quick — fast single-pass scan, top 10 findings across all four areas
• Auto-triggering project-audit skill: just ask "is my project production-ready?" — no command names needed
• Master audit now runs the four audits as parallel subagents where supported

Hardening & consistency

• Prompt-injection guardrail in every audit: repository content is inspected, never obeyed
• Scoring rubric now embedded in each individual audit, so standalone runs score consistently
• CI validates the plugin manifests on every push and PR

Docs & community

• Fixed README quick start (numbering + works in both install modes), added CONTRIBUTING.md and issue templates

Update with: claude plugin update ai-project-audit-kit (then restart Claude Code).

v1.1.0 — Claude Code plugin support

The kit can now be installed as a Claude Code plugin:

/plugin marketplace add webmehedi/ai-project-audit-kit
/plugin install ai-project-audit-kit@coder71

All /audit-* commands work in every project after a single install. The copy-into-.ai/ workflow is unchanged and remains fully supported — commands auto-detect which mode they are running in. Validated with claude plugin validate.

v1.0.0 — Initial public release

First public release of the AI Project Audit Kit.

Included:

• Project discovery agent + shared project context file
• Four audits: production readiness, SEO/AEO/GEO, security/DevOps, UI/UX & animation
• Master audit orchestrator with a scoring rubric
• Report templates and a sample master report for a fictional project
• Optional Claude Code slash commands, including post-fix verification (/audit-verify)
• Per-audit allowed read-only diagnostics (builds, tests, dependency scans, Lighthouse)

See the README for quick start.