spam-blocker

Systematic evidence-based debugging using runtime logs. Generates hypotheses, instruments code with NDJSON logs, guides reproduction, analyzes log evidence, and iterates until root cause is proven with cited log lines. Use when the user reports a bug, unexpected behavior, or asks to debug an issue.

Formats commit titles and descriptions using Conventional Commits. Use when suggesting or writing git commit messages.

Retrieves up-to-date docs for libraries and frameworks. Use when checking current Fastify, Vitest, Zod, or Node APIs before coding.

Removes overengineered or noisy AI-generated code from recent backend changes. Use when a diff reads as too defensive or too abstract.

Helps discover and install useful agent skills for this repo. Use when a user asks how to do a recurring task or asks whether a skill exists for Fastify, Vitest, Zod, security review, or similar backend workflows.

Resolves merge conflicts conservatively and verifies the repo still builds. Use when the tree contains conflict markers or git reports unresolved merges.

Orchestrates a multi-task implementation by delegating backend slices to plan-implementer subagents. Use when the user provides a plan and wants it executed in parallel where possible.

Formats concise GitHub issue titles and descriptions. Use when turning a problem into a tracking issue.

Automates browser verification for local routes and iframe flows. Use when checking challenge pages, OAuth status views, or other browser-visible server output.

Update or rewrite README.md for this backend service. Use when setup commands, workflow commands, API behavior, or architecture descriptions change.

Performs a cleanup pass on recent changes to simplify backend code without changing behavior. Use when the user asks for a refactor or cleanup.

Maintains the risk scoring system when factors, weights, or scoring behavior change. Use when working under packages/server/src/risk-score.