Custom Claude Code skills — multi-model AI delegation, structured git commits, code review, quick review, readability review, human review guide, codebase explanation, project handbook generation, codebase quality audit, devlog writing, retrospectives, interaction review, markdown export, and screenshot analysis
Based on adoption, maintenance, documentation, and repository signals. Not a security audit or endorsement.
Use when the user invokes /ai-council to consult three frontier AI models (Claude, GPT, Antigravity) in parallel and synthesize their responses into a consensus analysis
Use when the user invokes /antigravity to delegate a prompt to Google Antigravity CLI (agy), or /antigravity resume to continue a previous Antigravity session. SKIP if the user wants you to answer directly — this skill exists to consult Antigravity, not to substitute your own answer.
Use when the user invokes /claude to delegate a prompt to Claude Code CLI, or /claude resume to continue a previous Claude session. SKIP if the user wants you to answer directly — this skill exists to consult a separate Claude instance, not to substitute your own answer.
Use when the user invokes /codebase-audit to run a language-agnostic codebase quality audit measuring up to 12 quality criteria + development velocity with industry benchmarks, grading, and actionable recommendations.
Use when the user invokes /codereview to analyze code for correctness, quality, architecture, reliability, security, and performance violations with concrete before/after examples.
Custom Claude Code skills and Codex-installable skills.
/plugin marketplace add joesys/joesys-skills
/plugin install joesys-skills
Install the skills globally for Codex on this computer:
python scripts\install_codex_skills.py
The installer adapts the Claude Code source skills into Codex-compatible copies and writes them to:
%USERPROFILE%\.codex\skills\joesys-skills
If CODEX_HOME is set, the installer writes to:
%CODEX_HOME%\skills\joesys-skills
Restart Codex after installing so it reloads the skill list.
The generated Codex package is not checked in; codex/ is intentionally
gitignored. To update an existing Codex install after changing source skills,
rerun:
python scripts\install_codex_skills.py
To validate the adapter without installing:
pytest tests\test_codex_adapter.py -q
Multi-model consultation — ask one model or all three in parallel. Default model identifiers and CLI flags for the delegation skills (/claude, /codex, /antigravity, /ai-council) are defined in shared/model-defaults.md.
Consult three frontier AI models (Claude, GPT, Antigravity) in parallel on the same question. Automatically gathers relevant context, dispatches to all three, then synthesizes a structured analysis with consensus points, tensions, and a confidence matrix. Saves results by default.
/ai-council "Should we use PostgreSQL or MongoDB for our user data?"
/ai-council --no-save "Quick question about caching strategies"
/ai-council --path ./notes "Compare REST vs GraphQL for this use case"
Delegate prompts to Anthropic's Claude Code CLI for code analysis, refactoring, and automated editing. Critically evaluates Claude's output with special attention to shared blind spots (Claude evaluating Claude). Supports named session resume.
/claude "explain the auth flow in this repo"
/claude --model sonnet "quick analysis of this function"
/claude --permission-mode acceptEdits "fix the lint errors"
/claude --effort max "deep analysis of the architecture"
/claude --bare "analyze without plugins or hooks"
/claude resume "follow up on that"
/claude resume my-review "continue from that session"
Delegate prompts to OpenAI's Codex CLI for code analysis, refactoring, and automated editing. Critically evaluates Codex's output and supports session resume.
/codex "explain the auth flow in this repo"
/codex --model gpt-5.3 "analyze this function"
/codex --sandbox workspace-write "fix the lint errors"
/codex resume "follow up on that"
Delegate prompts to Google's Antigravity CLI (agy) for code analysis, refactoring, and automated editing. Critically evaluates Antigravity's output and supports session resume.
/antigravity "explain the auth flow in this repo"
/antigravity resume "follow up on that"
/antigravity resume <ID> "continue from that session"
Automated code analysis — from quick bug scans to full quality audits.
Dispatch 7 parallel domain-expert subagents to analyze code for violations across correctness, clean code, architecture, reliability, security, performance, and story readability. Produces a severity-grouped report (P0-P4) with concrete before/after fixes in the target language. Supports branch diffs, directory scans, single files, PR reviews, and commit reviews.
| Domain | Focus |
|---|---|
| Correctness | Actual bugs — wrong logic, off-by-one, null dereferences, race conditions |
| Clean Code | Naming, DRY (Rule of Three), nesting depth, function length, KISS/YAGNI |
| Architecture | Coupling, cohesion, layering violations, dependency direction, god classes |
| Reliability | Error handling, silent failures, resource leaks, missing validation |
| Security | Injection, hardcoded secrets, auth gaps, input sanitization, data exposure |
| Performance | Algorithmic complexity, N+1 queries, missing caching, unnecessary allocations |
| Story Readability | 8-dimension story-readability scoring (rolled into P2-P4 findings) |
/codereview # Review current branch diff vs. base
/codereview src/ # Scan all files in a directory
/codereview --file src/main.py # Review a single file
/codereview --pr 123 # Review files changed in a GitHub PR
/codereview --commit abc123 # Review files changed in a specific commit
/codereview --min-severity P1 # Only show P1+ findings (combinable with any mode)
Own this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimOwn this plugin?
Verify ownership to unlock analytics, metadata editing, and a verified badge. GitHub access is read-only (username + org membership).
Sign in to claimnpx claudepluginhub joesys/joesys-skills --plugin joesys-skillsPersonal Claude Code + Codex dev stack: security hooks, AI-first code conventions, /security-review, /repo-map, /stack-check, portable statusline. Designed to complement other skills-based plugins, not replace them.
Filip-PodstavecCodex CLIにセカンドオピニオンを求めるスキル。「codexと相談して」「codexに聞いて」で発動する。
shokaiLive codebase visualization and structural quality gate — 14 health dimensions graded A-F, dependency analysis, and architecture governance via MCP
sentruxAutomated code review with severity levels and actionable feedback
rohitg00Codebase exploration, refactoring, and quality analysis
UniswapComprehensive skill pack with 66 specialized skills for full-stack developers: 12 language experts (Python, TypeScript, Go, Rust, C++, Swift, Kotlin, C#, PHP, Java, SQL, JavaScript), 10 backend frameworks, 6 frontend/mobile, plus infrastructure, DevOps, security, and testing. Features progressive disclosure architecture for 50% faster loading.
Jeffallan