A collection of Agent Skills by Codeminer42 — reusable capabilities that help AI coding agents perform specialized tasks more effectively.
| Skill | Description |
|---|---|
| adr-setup | Set up Architecture Decision Records in a repository: ADR folder, template, index README, a first ADR, and agent instructions so agents check existing decisions and record new ones. Optionally adds a check (one bash script, also wired into package.json scripts in Node projects) and writes the job for the project's own CI, so CI fails on a stale index, dead links in agent instructions, citations of records that don't exist, and edits to merged records. |
| capability-map | Draft a capability map of the client's domain: one page with every capability's language, what it owns, its contracts, who acts, what it is not, where its code lives, its ADRs and state. Builds it from the client's documents as well as the code, keeps the client's words, records disagreements and ends with questions for the team and the client. Adds a generated capability table to the agent instructions, and optionally a bash check that fails CI on a stale table, a broken entry or a glossary term no capability speaks. |
| review-that-runs | Review a pull request by running it: detects the forge and stack, isolated worktree, the project's own CI gates, the app driven in Chrome for frontend changes, every finding verified and written to a local notes file, worktree and containers removed afterwards. Posts nothing to the code host. |
| sherlock | Autonomous investigation loop for finding and fixing performance issues, flaky tests, memory leaks, and reliability problems. |
| threat-model | Write a threat model kept in the repository: premise agreed with the team, assets, actors, trust boundaries and every known threat with a stable ID, severity, status, STRIDE class and code citations, plus the controls already in place and what is out of scope. Adds agent instructions so changes that cross a boundary cite the threats they touch. Inferred threats stay To verify; fixes nothing, accepts nothing, commits nothing. |
| privacy-assessment-rails | Assess a Rails app's full codebase for compliance with privacy laws, like GDPR and LGPD. |
| privacy-by-design-rails | Privacy-by-design patterns for Rails features that handle personal data — encryption, consent flows, DSAR endpoints, anonymization, and compliance with privacy laws like GDPR and LGPD. |
| privacy-review-rails | Review uncommitted or recently changed files for privacy-by-design rule violations (based on privacy laws like GDPR and LGPD) before committing. |
Install all skills from this repository:
npx skills add Codeminer42/skillsInstall a specific skill:
npx skills add Codeminer42/skills --skill skill-nameFor Claude Code, copy a skill directory directly:
cp -r skill-name/ ~/.claude/skills/Skills are folders of instructions, scripts, and resources that AI agents can discover and use to perform better at specific tasks. They follow the open Agent Skills specification and work across compatible agents including Claude Code, GitHub Copilot, Cursor, and others.
Each skill contains at minimum a SKILL.md file with YAML frontmatter (name and description) followed by markdown instructions.
See CONTRIBUTING.md for guidelines on creating and submitting new skills.
This project is licensed under the MIT License. See LICENSE for details.