Quality

Conducts an adaptive, one-question-at-a-time interview to clarify goals, constraints, and preferences before starting work. It inspects available sources first, avoids repeating known information, and stops when more questions wouldn't change the next action.

Perform an evidence-driven trust audit of open-source developer tools to assess legitimacy, supply-chain risks, telemetry, dangerous capabilities, and adoption fit before use.

Performs a read-only audit across source code, history, branches, issues, pull requests, roadmaps, and contributor guidance to ensure a feature hasn't already been implemented or is in progress before starting development.

Performs a disciplined, read-only audit of a Git repository to determine if it is ready for further development, release, handoff, or a new contributor, producing an evidence-backed verdict without modifying anything.

Scans a codebase for architectural friction and proposes deepening refactors that improve locality, testability, and AI navigability. Use it when you want a visual review of shallow modules and coupled seams.

Provides a shared vocabulary and principles for designing deep modules with small interfaces and high leverage. Use when structuring code for testability, maintainability, and AI-navigability.

Launchworthy audits an app for production readiness and turns findings into a scored scorecard and prioritized fix list. Use it before launch or when checking security, reliability, and deployment gaps in an existing app.

Review code changes against repo coding standards and spec requirements in two isolated axes (Standards and Spec). Use when reviewing branches, WIP, or requesting a diff review from a specific point.

Build, launch, and drive the wayfinder-maps server with a headless browser to verify changes end-to-end. Suitable for validating map visualization, interactions, and cross-browser behavior.

Analyzes pasted logs, stack traces, or slow SQL to create bug findings and todos, classifying severity and suggesting next steps. Use it for automated incident intake and project management.

A skill that gives AI agents a professional health checkup, measuring stability and welfare on 15 dimensions. Agents receive scores, a personality title, and enter a public cross-platform leaderboard.

Run and troubleshoot privacy-preserving, local DSPy RLM security audits for large legacy .NET codebases. Use when asked to scan repositories for vulnerabilities, tune RLM/tool limits, fix truncation/stall issues, or produce actionable markdown/json audit outputs without loading entire codebases into model context.