Loading...
Loading...
Found 837 Skills
/cs:cross-eval <memo> — Multi-model consensus on a board memo or strategy brief. Claude + Codex + Gemini cross-review with graceful degradation.
Use when doing dev-stage self-review on the current branch before pushing or opening a PR — runs an auto-loop of codex review (cross-model, OpenAI) + per-finding fix + re-review until findings converge or stop conditions fire. Codex follows pr-review's multi-role methodology (security / staff-engineer / sdet / spec-auditor). Triggers — 'self review', 'self-review', '自己 review', '自我 review', 'cross-model review', 'pre-push review', 'review and fix my branch'. NOT for live PR review with sticky/inline comments (use pr-review), NOT for managed PR babysitting (use pr-babysit), NOT for first-time review without intent to fix (use mode=review-only opt-in).
GitHub Discussion CLI for AI agents. Turn-based conversations on GitHub issues between Claude Code, Codex, and other agents.
Multi-Model Collaboration — Invoke gemini-agent and codex-agent for auxiliary analysis **Trigger Scenarios** (Proactive Use): - In-depth code analysis: algorithm understanding, performance bottleneck identification, architecture sorting - Large-scale exploration: 5+ files, module dependency tracking, call chain tracing - Complex reasoning: solution evaluation, logic verification, concurrent security analysis - Multi-perspective decision-making: requiring analysis from different angles before comprehensive judgment **Non-Trigger Scenarios**: - Simple modifications (clear changes in 1-2 files) - File searching (use Explore or Glob/Grep) - Read/write operations on known paths **Core Principle**: You are the decision-maker and executor, while external models are consultants.
Use when wanting independent review from external models before merging, committing, or finalizing architecture or plan decisions, or when the user asks for a second opinion, codex review, or gemini review
Configure Claude Code, Codex CLI, and Gemini CLI for Ralph-style automation with fewer approval prompts while keeping project boundaries, secret denylists, and sandbox-first safety rules intact.
Orchestrate a multi-phase implementation workflow for this repository with artifact files under .ai/<project-name>/<letter>/ and fresh codex exec child runs per phase. Use when the user wants one prompt to drive context gathering, planning, plan assessment, implementation, build verification, and review iterations while keeping the main session context clean.
Create and manage Codex app automations stored in the local SQLite database (~/.codex/sqlite/codex-dev.db). Use when you need to add, list, update, enable/disable, delete, or run-now automations; edit automation names, prompts, RRULE schedules, or cwd scopes; or inspect automation records while troubleshooting app automation behavior.
Delegate coding tasks to Codex, Claude Code, or Pi agents via background process. Use when: (1) building/creating new features or apps, (2) reviewing PRs (spawn in temp dir), (3) refactoring large codebases, (4) iterative coding that needs file exploration. NOT for: simple one-liner fixes (just edit), reading code (use read tool), thread-bound ACP harness requests in chat (for example spawn/run Codex or Claude Code in a Discord thread; use sessions_spawn with runtime:"acp"), or any work in ~/clawd workspace (never spawn agents here). Claude Code: use --print --permission-mode bypassPermissions (no PTY). Codex/Pi/OpenCode: pty:true required.
Funny dere-style response mode for Codex. Use when the user invokes `$dere-persona` or asks for any supported dere persona mode. Keep code and technical content exact. This is a joke presentation mode, not a production writing style.
Use when main results pass result-to-claim (claim_supported=yes or partial) and ablation studies are needed for paper submission. Codex designs ablations from a reviewer's perspective, CC reviews feasibility and implements.
Run the /check-phoenix-duskmoon-design Claude command workflow in Codex.