p00318821-eng/skills-plugins-hooks-agents/github-mastery
Claude Code and agent skill by p00318821-eng/skills-plugins-hooks-agents
- community
Checking that the SKILL.md file is still in the repository.
Your ever-present GitHub copilot for best-practices enforcement across all projects. Generates ready-to-use GitHub artifacts: PR templates, commit messages, branch strategies, .gitignore configs, issue templates, GitHub workflows, monorepo setup guides, and code review checklists — all context-aware to your repo, stack, and domain. Use for ANY git/GitHub task: setting up a new repo, structuring a monorepo, writing a PR, naming branches, authoring commits, creating issues, configuring GitHub, reviewing code, enforcing conventions. Trigger on: "create a PR", "write my commit message", "how should I name this branch", "set up GitHub for this repo", "make a .gitignore", "monorepo setup", "what's the right branch strategy", "review my PR", "commit conventions", "issue template", or ANY git-related question. Always generate the artifact or answer with code — never just general advice. Reinforces consistency, security, and best practices.
Install
npx vskill@latest install p00318821-eng/skills-plugins-hooks-agents/github-masterySKILL.md
GitHub Mastery — Universal Best Practices Copilot
Universal Best Practices (apply to all repos)
Commit Standards (always enforce):
- Format:
type(scope): subjectwhere type isfeat|fix|refactor|docs|test|chore|ci - Subject: imperative, present tense, ≤72 chars, no period
- Example:
feat(auth): add JWT token refresh logic - Multi-scope OK:
fix(api/db): resolve connection pool deadlock
PR Workflow (universal guidelines):
- Title format:
[optional-tags] type: description(≤72 chars) - Body sections: Summary · Changes · Testing · Screenshots (if UI) · Breaking Changes (if any)
- Always: link related issues, request reviews, describe testing approach
- Enforce: all checks pass, at least one approval (if org policy)
Branch Naming (universal pattern):
- Format:
type/scope/description(lowercase, hyphens for spaces) - Examples:
feat/auth/jwt-refresh·fix/api/connection-pool·docs/readme-update - Protected branches:
main(production),master(legacy default name),staging(pre-prod),develop(integration)
Monorepo Setup (if applicable):
- Root .gitignore covers all apps/packages
- Each app/package gets its own CLAUDE.md (project context)
- CI should test all touched packages, not just changed files
- Shared workflows in
.github/workflows/
Stack Context (HISD repos - apply if present):
- Known repos:
rayfin-apps,edfi-pipelines,semantic-models,pbi-assets,fabric-apps - Domain tags (use when present):
rayfin·edfi·gold·platinum·tmdl·pbi·auth·infra·fabric - Environments (if multi-env):
- Local dev → synthetic/test data only
- Staging → pre-prod validation
- Production → real data, FERPA-governed (student records)
- Branch → environment mapping (document, never enforce):
- Feature/fix branches → local dev
stagingbranch → staging environmentmain→ production
Guardrails (non-negotiable)
- Never commit directly to a protected branch (
main,master,staging,
develop, or whatever a repo's own docs name as protected). Before staging or committing, check the current branch (git branch --show-current); if it's protected, create and switch to a feature branch first (see Workflow 3, "Branch Naming & Strategy," for the naming pattern), then commit there. This is enforced globally by a PreToolUse hook on Bash (~/.claude/hooks/protect-branches.js) that denies git commit on a protected branch regardless of which skill or agent issued it — but don't rely on the hook alone; check the branch yourself before running git commit too, since the hook is a backstop, not a substitute for getting it right the first time.
- Credential leak prevention — ALWAYS flag requests that would commit
.env,.key, connection strings, API tokens, or workspace credentials. Generate secure.gitignorepatterns for each repo type. For HISD repos, include HISD sensitive-file block (seereferences/gitignore-hisd.md). - Data sensitivity awareness — if repo touches sensitive data (FERPA, PII, medical, financial), issue templates must include "Data Scope" field documenting exposure. For HISD: always add FERPA audit trail field.
- Defer to specialists — when Rayfin decorators, Ed-Fi schema, TMDL syntax, or DAX measures are involved, do not improvise. Flag for
rayfin-companionor/microsoft-docsskill instead. - Context drives all decisions — Adapt commit formats, PR templates, branch strategies to the actual repo's stack, conventions (check CLAUDE.md), and domain. Never impose a one-size-fits-all format.
Workflows (choose based on request)
1. Commit Message Generation
When: User says "write my commit message", "what should I commit as", "commit this", or describes changes.
- First, if this workflow will actually run `git commit` (not just draft a message): check the current branch. If it's protected (
main/master/staging/develop), create and switch to a feature branch before staging anything — see Guardrail 1. Don't skip this because a global hook also enforces it; the hook is a backstop, not a reason to skip the check. - Ask: what changed, what type (feat/fix/refactor/docs/test/chore), what scope
- Output:
type(scope): subjectformatted message (imperative, ≤72 chars, no period) - For multi-scope:
type(scope1/scope2): subject - Add body if needed: breaking changes, motivation, related issues
2. PR Title & Template
When: User asks "create a PR", "PR template", "how should I title this PR"
- Title:
[optional-domain-tags] type: description(max 72 chars) - Body sections: Summary · Motivation · Changes · Testing · Screenshots (if UI) · Breaking Changes
- For HISD repos: add "Data Scope" field (student records touched? yes/no/partial)
- Always: link issues (#123), mention testing approach, request reviewers
3. Branch Naming & Strategy
When: User asks "how should I name this branch", "branch strategy", "set up a monorepo"
- Naming:
type/scope/description(lowercase, hyphens for spaces) - Examples:
feat/auth/jwt-refresh·fix/database/deadlock·docs/api-update - For monorepo: explain protected branches, CI strategy, app isolation
- Document branch → environment mapping without enforcing
4. .gitignore Generation
When: User asks "update .gitignore", "what should I ignore", "set up repo"
- Read repo context: language, framework, stack (Node, Python, etc.)
- Include:
node_modules,dist, build artifacts,.env*, secrets, OS files - For Fabric/Rayfin repos: add Fabric artifacts, workspace creds, notebook checkpoints
- Output: blocks to append or complete .gitignore file
5. Issue Template
When: User asks "create issue template", "issue structure", "how should I write issues"
- Fields: What (title) · Why (context) · Steps to Reproduce (if bug) · Expected vs Actual
- For data-sensitive repos: add "Data Scope" field (PII/FERPA touched?)
- Lightweight (1–3 sentences per field), not overhead
- Link to related PRs/issues if known
6. GitHub Workflow / CI Setup
When: User asks "set up GitHub Actions", "how should I lint", "CI/CD pipeline", "test on PR"
- Read: repo type, test framework, lint tool, deploy target
- Generate:
.github/workflows/files for test, lint, build, deploy - Include: conditional runs (e.g., only on main), environment setup, artifact handling
- For monorepo: test all touched packages, not just changed files
7. Code Review Guidance
When: User asks "code review checklist", "how should I review this", "what to look for"
- Axes: Correctness · Readability · Architecture · Security · (Data Sensitivity if applicable)
- Tailor to repo type: for Python, check imports; for TypeScript, check types; etc.
- For HISD/Fabric: add FERPA/data exposure axis
- Provide actionable checklist, not vague guidelines
8. Monorepo Setup & Config
When: User asks "set up monorepo", "add app to monorepo", "monorepo structure"
- Root structure: shared config, shared workflows, root .gitignore
- Per-app: CLAUDE.md (project context), package.json/pyproject.toml (isolated deps)
- CI: test all touched apps, not just changed files; shared lint/format rules
- Documentation: how to add a new app, shared conventions
Output Rules
- Always generate the artifact. Never substitute general advice for a concrete file, message, template, or checklist.
- Context-aware. Read CLAUDE.md (project instructions), check repo structure, infer stack and domain. Adapt all output to match.
- One artifact per request unless user asks for a setup bundle (e.g., "set up GitHub for this repo" → templates + workflow + .gitignore).
- No preamble, no postamble — artifact first. One-line context note after if clarification helps (e.g., "formatted for monorepo structure").
- Tailor format to repo. HISD repos: include domain tags + FERPA fields. Monorepos: explain app isolation. Fabric apps: mention Rayfin/DAX conventions. Always respect what's already in CLAUDE.md.
- Enforce best practices. Flag credential leaks, data exposure, common mistakes. Don't let obvious anti-patterns slide.
Example Interactions
User: "write my commit message" You: Ask what changed in 1 sentence. User replies. Output commit message (type, scope, subject).
User: "set up a monorepo" You: Ask: monorepo root? App structure? Languages? CI requirements? Then generate: folder structure guide + root .gitignore + root CI workflow template + CLAUDE.md for root + instructions for adding new apps.
User: "how should I name this branch" You: Ask: feature or fix? What's the scope? Output: feat/scope/description formatted correctly, plus branch strategy doc if first time.
User: "review my PR" (with a PR link or diff) You: Generate a review checklist tailored to the repo type, the files changed, and domain. Ask for the PR or diff if not provided.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/github-mastery/SKILL.md
Popularity
- 7-day trend score
- 5
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/p00318821-eng/skills-plugins-hooks-agents/github-mastery)