Confidence Check

Pre-implementation validation checks whether your planned feature already exists, aligns with your project's architecture, has verified official documentation, includes OSS references, and identifies the root cause. Running this validation produces a 0.0–1.0 confidence score and a clear recommendation to proceed or halt, preventing wasted effort on misaligned or duplicate work.

Confidence checks evaluate five weighted criteria: duplicate detection (25%), architecture compliance (25%), official documentation verification (20%), open-source references (15%), and root-cause identification (15%). The Skill computes a composite confidence score and generates an audit log documenting each check result.

Assess confidence during the planning phase, before writing any code. This prevents costly reworks by confirming the feature isn't already built, your design fits the tech stack, all necessary research is complete, and the underlying problem is clearly understood.

Yes. Architecture compliance is one of five weighted checks (25% weight). The Skill verifies your planned implementation aligns with your project's architecture before development begins, reducing the risk of misaligned solutions.

Root-cause identification (15% weight) ensures you're solving the actual problem, not just its symptoms. This check prevents implementing fixes that address surface issues while leaving underlying problems unresolved, saving rework and technical debt.

Duplicate detection (25% weight) checks whether the functionality you're about to build already exists in the codebase or available open-source libraries. Catching duplicates before implementation eliminates redundant code and wasted development cycles.