lousy-agents
lousy-agents 收录了来自 zpratt 的 8 个 skills,并提供仓库级职业覆盖和站内 skill 详情页。
这个仓库中的 skills
Rewrites the commits on a pull request's head branch so the changelog and release notes semantic-release generates read as a coherent story. Use when asked to 'curate release', 'curate commit history', 'rewrite commits before merge', 'rebase this branch into clean commits', 'fix commit messages for release notes', 'clean up the commits on this branch', or when an unattended routine must reshape a PR's commits into Conventional Commits. Not for changing file contents, squashing at merge time, or writing one commit message by hand.
Use when the user asks to turn a feature request, idea, PRD draft, or backlog issue into a structured EARS-format spec. Trigger phrases include "draft a spec", "plan a feature", "scaffold a spec", "write a feature spec", "convert this issue to a spec", "plan this issue", or invocation via /feature-to-plan. Do NOT use for reviewing an existing spec PR (use triaging-pr-reviews) or for editing a specific section of an already-drafted spec (use a direct Edit instead).
Uncover test coverage gaps by applying semantic mutations to production TypeScript code and identifying which mutations survive (tests still pass). Surviving mutations indicate areas where tests are insufficient to detect behavioral changes.
Converts an approved local spec, master plan, or GitHub epic issue into a GitHub Issue dependency graph with native sub-issues and blocking relationships. Use when asked to 'convert plan to issues', 'create GitHub sub-issues', 'populate issues from a spec', 'plan to graph', or 'break down a GitHub epic into tasks'.
Generates adversarial, negative, and chaos tests for TypeScript code. Use when asked for evil tests, security tests, boundary tests, fuzzing, injection testing, rugged testing, or to 'break this' / 'find vulnerabilities'. Not for happy-path or standard unit/integration tests.
Adversarially review feature specifications, implementation plans, GitHub issues, PRDs, or EARS-format specs before coding. Use when the user asks to audit, critique, stress-test, validate, review for ambiguity, find contradictions, identify gaps, harden acceptance criteria, prepare a spec for Codex, GitHub Copilot, Claude, or another coding agent, or produce a structured flaw list for a spec-improvement loop.
Use when triaging or analyzing PR review comments — especially from automated reviewers like GitHub Copilot — to classify root concerns, verify claims against actual code, evaluate trade-offs, and decide what to implement, reject, or implement differently. Also use to process or respond to review feedback, handle Copilot suggestions, or sort through code review comments
Use when the user asks to turn a feature request, idea, PRD draft, or backlog issue into a structured EARS-format spec. Trigger phrases include "draft a spec", "plan a feature", "scaffold a spec", "write a feature spec", "convert this issue to a spec", "plan this issue", or invocation via /feature-to-plan. Do NOT use for reviewing an existing spec PR (use triaging-pr-reviews) or for editing a specific section of an already-drafted spec (use a direct Edit instead).