v0.2.0 strengthens Agent Readiness Scanner as deterministic agent preflight infrastructure.
What's new
- Critical-failure visibility for committed .env files
- Critical-failure visibility for hardcoded secret-pattern findings
- JSON, terminal, and Markdown outputs now surface critical failures separately from score
- Score formula unchanged
- README positioning sharpened around Claude, Cursor, Copilot, Codex, and local agents
- New roadmap and v0.2 build report
Why it matters
A repo can no longer appear mostly ready while hiding high-severity blockers. Agents are powerful, but they need a runway. This tool checks the runway.
Verification
- 197 tests passing
- Self-scan: 100/100 GREEN
- Fixture validation: 5/5
- Fixture 4 remains YELLOW but now reports 2 critical failures
Still intentionally not included
- SaaS
- dashboard
- billing
- auth
- telemetry
- LLM calls
- Token Burn Firewall
- Repo Red Cell Bot