Find the perfect skill

Testing

Run a manual eval loop for the brepjs-cad implement skill: re-author playground examples, have a blind judge grade them, and propose SKILL.md fixes. Includes verify mode with a precision/recall harness.

Testing

Audit the .spt TLV walker and placeholder-billboard fallback, focusing on byte accounting, fallback correctness, and cross-cutting wiring.

Testing

Establishes baselines for page load times, Core Web Vitals, and resource sizes. Compares before/after on each PR and tracks performance trends.

Testing

Audit NES APU register writes for hardware correctness across Pulse, Triangle, Noise, DPCM channels, verifying code against documented hardware semantics.

Testing

Live developer experience audit: tests docs, measures onboarding time, and produces a DX scorecard with evidence.

Testing

Technical audit of ByroRedux's Fallout 4 support: NIF, BA2, BGSM, M49 precombines. Confirms shipped features still hold and flags regressions.

Testing

Master comprehensive testing strategies for smart contracts using Hardhat, Foundry, and advanced testing patterns.

Testing

Coordinates naive-user comprehension prompts and logs results to a dated findings report. Invoke once per scenario, paste response manually.

Testing

Launches a separate agent to objectively review skill files against Anthropic best practices, scoring 100 points across 5 axes.

Testing

Turn settled intent into the cheapest discriminating executable proof before implementation or candidate promotion. Uses spec-to-test and review-tests to strengthen and validate proof.

Testing

Creates a draft from facts, norms, evidence, and motion; provides a deadline and risk traffic light with immediate steps.

Testing

Provides step-by-step guidance for test retry configuration, covering best practices for unit, integration, and mocking with frameworks like Jest and Pytest.