claude-code-plugins

Invoke OpenAI Codex CLI for cross-model collaboration

Claude Code only. Use when the user explicitly invokes /agents-bridge:codex-orchestrate (or /codex-orchestrate) in Claude Code or asks Claude to orchestrate a large multi-slice plan by delegating implementation to the Codex CLI through agents-bridge. In native Codex, use $slice-runner instead.

Critical second opinion from Codex (OpenAI) on a proposal Claude just made — confirms what is sound, challenges what is genuinely weak, and suggests a better path with clear reasoning when one exists. Constructive, not contrarian. Use to cross-check a Claude design, refactor, API, or fix with a non-Claude model before acting on it.

Start a session in pair-planning mode where Claude and Codex (read-only) each draft an independent implementation plan for a task, then cross-review through an open-point ledger until consensus, escalating remaining disagreements to the user. Produces one agreed plan; does not write code. Use at the start of a task to align two frontier models on the approach before implementation.

Audits CLAUDE.md and AGENTS.md files against research-backed best practices: instruction budget scoring, anti-pattern detection, staleness checks, and concrete fix proposals. Use when the user asks to audit, check, review, or optimize context files, or mentions "instruction budget" or "context file quality". Treats CLAUDE.md and AGENTS.md identically.

Synchronizes project CLAUDE.md with recent codebase changes by analyzing git history, reviewing against official Anthropic best practices with parallel agents, and proposing updates for user approval. Use when CLAUDE.md is outdated, missing, or drifted after major architectural changes.

Explains complex concepts -- technical, scientific, or philosophical -- to intelligent non-specialists using mechanism-mapping analogies, progressive depth, and honest simplification markers. Use when the user asks to demystify, break down, or explain a concept for a general audience: "demystify X", "what is X", "how does X work", "explain like", "ELI-smart". NOT for codebase-specific explanations (use /explain for that).

Show plugin-dev status and explain sync workflow

Convert rough coding, product, architecture, audit, review, migration, cleanup, debugging, or orchestration intent into the smallest useful goal payload to hand to a fresh Claude Code agent (a new session, a subagent, or /loop). Invoke with /goalify, or "/goalify interactive" to be questioned one point at a time before the goal is written. Produces goal text only, never the downstream implementation.

Create opinionated, anti-slop skills that encode taste. NOT for: general skill structure, reference guides, procedural docs (use skill-creator instead). USE for: frontend-design, cli-design, system-architecture style skills. Triggers: "anti-slop skill", "opinionated skill", "make Claude less generic", "skill like frontend-design", "design-excellence skill".

Iteratively improve artifacts via parallel reviewers with anchor-based drift prevention. NOT for: one-shot edits, simple fixes, tasks without measurable quality. USE for: skills, prompts, commands, agents needing iterative refinement with score targets. Triggers: "iterate to 9.0", "quality loop", "IAQI", "reviewer loop", "improve until threshold", "iterative quality".

Audit an Effect codebase for silent error swallowing -- catchAll patterns that collapse errors into sentinel values (null, [], false, void) in the success channel. Runs a scanner script, spawns a read-only auditor, and produces a classified report. NOT for: non-Effect projects, React/frontend code, plain TypeScript without Effect.

Audit a codebase for handcrafted code that duplicates functionality already available in the project's dependencies. Reads package.json, launches parallel exploration agents, verifies replacement feasibility, and produces a structured refactor plan. Audit only -- does not execute changes.

Adaptive pre-planning exploration loop. Builds sufficient context before committing to a detailed plan. Use when: (1) brief is ambiguous or underspecified, (2) task touches unfamiliar code or libraries, (3) multiple valid approaches exist and the right one isn't obvious, (4) user says "thorough plan" or invokes /thorough-plan. NOT for trivial tasks where the path is obvious — triage will skip exploration in those cases.

Ship a feature branch: prepare a clean PR (strip working files with intelligent suggestions), create/update the GitHub PR, then merge to main with GPG signing when ready. Detects lifecycle stage automatically. Use when: "ship", "prep pr", "create pr", "open pr", "merge to main", "submit for review", "push for review", "ship this branch".

Clean stale git branches and worktrees interactively. Audits local/remote branches, detects squash-merged and orphaned branches, presents categorized cleanup with user confirmation. Use when: "git sweep", "sweep branches", "branch cleanup", "prune stale branches".

Verify no behavioral regressions after implementation. Spawns parallel sonnet subagents for semantic diff review + test execution. Use when: user says "verify no regressions", "check for regressions", "make sure nothing broke".

Rigorous planning workflow with research validation, task breakdown, dependency mapping, and post-implementation verification. Use when: (1) planning complex features touching 5+ files, (2) architectural changes or major refactors, (3) unfamiliar tech requiring doc research, (4) user says "thorough plan", "rigorous plan", or "structured plan". Invoke with /structured-plan before or during plan mode.

Validate plugin structure, versions, and marketplace sync. Use when running CI checks, gating PRs, verifying marketplace.json sync, or auditing plugin directories for version consistency and required fields.

Thorough planning for complex features using Task-based orchestration. Turn discussions into executable, self-contained implementation plans. USE for multi-file features, architectural changes, unfamiliar tech requiring research. SKIP for quick fixes or single-file changes.

Generate comprehensive tests for architectural layers with coverage-first analysis. Use when testing specific layers (core, domain, application, infrastructure, boundary). Reads testing strategy from playbook or uses interactive template selection.

Create distinctive, user-respecting command-line interfaces with exceptional UX. Use this skill when the user asks to build CLI tools, scripts, or terminal applications. Generates well-crafted CLIs that follow platform conventions and treat users as humans, not machines.

Design systems with appropriate complexity—no more, no less. Use this skill when the user asks to architect applications, design system boundaries, or make structural decisions. Generates pragmatic architectures that solve real problems without premature abstraction.

This skill should be used when the user asks to "set up submodules", "migrate branches to submodules", "automate submodule sync", "create submodule architecture", "convert branches to repos", or mentions setting up git submodules with GitHub Actions automation for multi-repo synchronization.

Create Mermaid diagrams. ONLY use when user explicitly says "Mermaid". NOT for general diagrams, schemas, ASCII art, or wireframes.