Quality

Agent Quality Critique Dimensions provides checks for Claude Code agent definitions, including template compliance, safety, examples, loading, and priority. Use it when reviewing an agent file for correctness and fit.

Agent Quality Critique Dimensions provides review criteria for Claude Code agent definitions, including template compliance, safety, examples, and loading behavior. Use it when validating an agent before release or refactor.

Agent Testing Framework defines a 5-layer approach for validating agent outputs, handoffs, and adversarial cases. It is used to check prompt injection resistance and other security constraints in Claude Code agents.

AT Completeness Check gates acceptance-test sets against a canonical 7-category taxonomy and 15-item checklist. Use it to spot missing boundary, state, lifecycle, flag, negative, and environment cases.

BDD Methodology defines acceptance test design with Given-When-Then scenarios, scenario-writing rules, and pytest-bdd implementation. Use it to write living documentation and avoid common BDD anti-patterns.

BDD Requirements Discovery guides collaborative discovery of requirements through Example Mapping, Three Amigos conversations, and concrete examples. Use it to turn business rules into Given-When-Then scenarios and uncover edge cases.

Canary detects whether auto-injection is active by exposing a known marker for verification. Use it to confirm skill-loading behavior during testing.

Nw Bug Fix guides a defect through root-cause analysis, user review, and a regression test plus minimal fix. Use it for known bugs that need a confirmed diagnosis before implementation.

Project Reading helps a buddy agent inspect a codebase in a defined order and answer questions with file-and-line citations. Use it when you need grounded project facts, architecture context, or current state without guessing.

E2E Coding helps with code generation, multi-language support, and code reviews for end-to-end testing workflows. Use it when an agent needs assistance writing or reviewing application code.

E2E Test is a test skill for end-to-end validation that returns structured JSON with a status field. Use it when you need basic query operations and consistent machine-readable responses.

PR Review performs a thorough local code review with full project context. Use it to inspect pull request changes before merging or when running /pr-review.