plug-n-skills

Use when locating, summarizing, auditing, or debugging Codex session rollout JSONL logs by CODEX_THREAD_ID, cwd, query, issue key, project path, child or inherited-history symptoms, malformed or huge log symptoms, permission concerns, or "what happened in this Codex thread" questions.

Use when auditing agent skills or plugins for safety, coverage, duplicated logic, token/context cost, prompt-contract risk, validation readiness, marketplace visibility, dependencies, network behavior, or install/update risk. For structural split/merge/move/delete/router decisions across skills or plugins, use capability-portfolio-architect instead.

Use when auditing agent skill/plugin portfolio architecture for cross-plugin capability overlap, weak routing, duplicated or missing skills, split/merge/delete/move decisions, shared capability extraction, reference/script extraction, or plugin boundary redesign. For safety, coverage, or validation review of a single artifact without boundary changes, use capability-auditor instead.

Use when performing broad external-first discovery, vetting, scoring, distillation, or synthesis of agent skills or plugin packs from the public web, GitHub/public repositories, OpenClaw/ClawHub, marketplaces, research sources, community implementations, user references, or local skills. Use for well-vetted capability synthesis, cross-skill comparison, plugin-pack synthesis, global capability acquisition, skill strengthening, and adoption/rejection changelogs.

Route and orchestrate agent skill and plugin lifecycle work. Use when work involves agent skills, plugins, marketplace packages, capability acquisition, synthesis, creation, installation, packaging, audit, or context-density optimization.

Use when creating, updating, validating, optionally installing/cache-refreshing, or handing off marketplace-backed agent plugins. Use for plugin manifests, local marketplace entries, plugin-pack synthesis, triggerable skill bundles, install visibility gates, and Codex app deeplinks.

Use when creating, refactoring, splitting, compressing, validating, or packaging agent skills. Use for SKILL.md trigger design, agents/openai.yaml metadata, skill resources, progressive disclosure, token-efficient instructions, quick validation, and plugin-contained skills.

Use when designing, auditing, refactoring, measuring, or validating token-efficient context and prompt/output contracts. Trigger for AGENTS.md, prompts, skill packages, marketplace plugins, MCP/tool schemas, long-context placement, prompt/context compression, retrieval provenance, runtime context diagnostics, duplicated or overlapping prose, competing skill descriptions, irrelevant-context pruning, prompt reformatting, agent/subagent context handoffs, brittle trigger descriptions, strict JSON/schema/tool-call output, validators, retry/repair loops, or behavior-preserving token reduction.

Create, refresh, audit, or migrate repository agent guidance files such as AGENTS.md, AGENTS.override.md, CLAUDE.md, .claude/rules, and Cursor rules. Use when a repo needs durable coding-agent instructions, instruction load-order decisions, nested guidance, or cleanup of bloated or stale agent docs. Do not use for ordinary human README or CONTRIBUTING docs unless they must feed agent guidance.

Find, list, vet, install, or update agent skills from curated catalogs, GitHub repo paths, local folders, or user-provided references. Use when installation must be preceded by provenance, safety, dependency, and capability checks.

Use when intended architecture must be compared with implementation: dependency policies, ADRs, ownership rules, drift, erosion, recovered models, and conformance classifications.

Use when architecture choices need executable guardrails: dependency rules, cycle checks, boundary tests, ownership review gates, ADR compliance, runtime/resilience checks, and CI fitness functions.

Use whenever code work has structural consequences: project architecture, module boundaries, dependency direction, runtime topology, ownership, ADRs, fitness functions, or architecture refactoring.

Use when architecture crosses ownership or review boundaries: CODEOWNERS/OWNERS, module coverage, cross-owned dependencies, socio-technical coordination, and governance paths.

Use when code work touches runtime shape: services, app/CLI/background flows, deployment/IaC, observability, resilience, external integrations, ownership, and runtime coupling.

Use before architecture-significant code work to recover the actual source-backed system shape: modules, dependencies, domain seams, runtime coupling, ownership, quality attributes, tests, docs, and risks.

Repair stale or false agent capability instructions immediately. Use when a skill, script, plugin manifest/helper, MCP/tool contract, connector guidance, validator, or bundled reference describes commands, schemas, paths, outputs, install state, dependencies, or behavior that does not match live reality.

Use when creating, editing, auditing, or debugging skill names, descriptions, SKILL.md frontmatter, plugin skill metadata, discovery triggers, routing phrases, under-triggering, over-triggering, or cases where a relevant skill is not being invoked or read.

Use when coordinating Claude Code background agents, `claude agents --json`, dispatched-session defaults, git worktrees, tmux/iTerm panes, resume/continue/from-pr/fork-session, session names and IDs, remote control, prompt suggestions, or cloud ultrareview runs.

Route Claude Code CLI work across local CLI inspection, interactive and print-mode automation, plugin and MCP lifecycle, diagnostics, hooks/settings, background agents, worktrees, sessions, remote control, and ultrareview.

Use when diagnosing Claude Code install/update/auth/config/runtime health, broken customizations, safe mode, bare mode, debug logs, auto-mode classifier, settings validation, keychain or API-key auth boundaries, IDE/Chrome integration, doctor warnings, or update/install/setup-token issues.

Use when creating, editing, auditing, or debugging Claude Code settings, hooks, CLAUDE.md/rules loading, custom agents, tool allow/deny permissions, output styles, workflows, setting sources, safe-mode/bare-mode differences, or plugin-provided customizations.

Use when managing Claude Code plugins, plugin marketplaces, plugin validation/details/token cost/tag/install/update/remove/prune, session-only plugin-dir or plugin-url, MCP list/get/add/remove/import/serve, project MCP approvals, strict MCP config, transports, headers, OAuth, or MCP/plugin installation failures.

Use when preparing, running, or debugging non-interactive Claude Code with `claude --print`, output-format text/json/stream-json, input-format text/stream-json, JSON schema validation, budget caps, fallback model, no session persistence, prompt suggestions, or trusted-directory CI-style runs.

Route Codex CLI work across local CLI inspection, non-interactive exec/review automation, plugin and MCP management, doctor/debug/sandbox/app-server diagnostics, session log forensics, and Codex app local environment actions.

Use when diagnosing Codex CLI install, config, auth, runtime, feature flags, sandbox denials, debug models, prompt input, app-server, remote-control, remote websocket/unix connections, or local Codex health failures.

Create, update, debug, or remove Codex app local project environments and actions, including `.codex/environments/environment.toml`, Run/Test/Preview buttons, startup commands, simulator/browser/dev-server launch actions, environment variables, and repeatable local commands.

Use when preparing, running, debugging, or reviewing non-interactive Codex CLI automation with `codex exec`, `codex exec resume`, `codex review`, JSONL events, output schemas, last-message files, cwd/profile/config flags, sandbox and approval modes, or CI-style agent runs.

Use when managing Codex CLI plugins, plugin marketplaces, local marketplace refreshes, cache visibility, MCP server list/get/add/remove/login/logout, MCP bearer token environment variables, or plugin/MCP installation failures.

Diagnose and fix Kotlin Multiplatform Gradle, source-set, dependency, Android target, Compose plugin, KGP/AGP, testing, static-analysis, and CI issues.

Review Kotlin Multiplatform production readiness, build governance, convention plugins, version catalogs, repository policy, module API hygiene, Klibs target support, ABI validation, publishing, and adoption risk.

Audit Kotlin Multiplatform production readiness with explicit scorecards, release blockers, risk ownership, validation commands, and deferred checks across architecture, build, testing, interop, security, performance, and publishing.

Route and execute Kotlin Multiplatform tasks across architecture, Gradle diagnosis, Compose Multiplatform UI, Android-KMP migration, iOS interop, CocoaPods or SwiftPM migration, testing, performance, security, CI, publishing, and production readiness.

Use for scientific or scholarly research with source traceability, literature reviews, paper discovery, arXiv/OpenAlex/Crossref/Europe PMC/Semantic Scholar/PubMed queries, corpus building, DOI deduplication, source-backed claim extraction, evidence synthesis, or research quality validation.

Audit SDD artifacts, requirement-task traceability, surface selection, and completion evidence before implementation or final delivery.

Execute SDD task lists safely, update task status, handle spec drift, and verify completion with fresh evidence before claiming done.

Convert approved SDD specs into technical plans, design artifacts, contracts, quickstart validation, and traceable task lists.

Route Spec-Driven Development work across lightweight, Spec Kit, Kiro-style, OpenSpec-style, brownfield, bugfix, planning, implementation, and audit lanes.

Use GitHub Spec Kit projects and commands, including constitution, specify, clarify, plan, tasks, analyze, implement, extensions, and presets.

Configure or review Tauri 2 tauri.conf files, capabilities, permissions, CSP, scoped filesystem/network/shell access, window labels, plugin permissions, and security-sensitive frontend API exposure.