{"id":"artemiimillier-bulletproof","name":"bulletproof","af_score":45.8,"security_score":60.0,"reliability_score":33.8,"what_it_does":"Bulletproof provides a 12-stage, spec-and-verification-first workflow (with templates and sub-agents) for using AI coding agents more safely and reliably—aimed at reducing regressions, ensuring acceptance criteria are met, performing impact analysis, running security scanning, and enforcing anti-rationalization gates before completion.","best_when":"When you want an AI coding workflow that enforces contracts (spec/acceptance criteria), verification steps, and gated completion to reduce “done when it isn’t done” outcomes.","avoid_when":"When you can’t run the required tests/verification steps (or don’t have a suitable test suite), since the workflow’s value depends on those gates.","last_evaluated":"2026-03-30T13:49:57.569142+00:00","has_mcp":false,"has_api":false,"auth_methods":[],"has_free_tier":false,"known_gotchas":["The workflow’s effectiveness depends on having runnable tests and being able to perform verification/impact analysis steps in your repo.","If the agent client or skill integration doesn’t support the described Agent Skills hooks/templates, the workflow may not be applied as intended.","AI-generated security scanning results still require interpretation/review appropriate for your risk tolerance."],"error_quality":0.0}