Skills de Claude Code · página 37

How to work in a Worldbuilding project (the `worldbuilding` starter pack): a fiction encyclopedia of characters, settings, themes, factions, and lore. Read when the project has these folders, OR when asked to add a character, setting, faction, or piece of lore, or to check the world for internal consistency. Carries the auto-stub and consistency behaviors so that guidance does not live inside template bodies or folder descriptions. Complements the platform `open-knowledge` skill; does not replace it.

How to work in a Writing Pipeline project (the `writing-pipeline` starter pack): a three-stage drafting flow, ideas → drafts → published. Read when the project has these folders, OR when asked to capture an idea, start or revise a draft, review a piece before it ships, or publish it. Carries the stage flow and review behaviors so that guidance does not live inside template bodies or folder descriptions. Complements the platform `open-knowledge` skill; does not replace it.

Authoritative agent-runtime contract for working inside an OpenKnowledge project — a markdown-CRDT knowledge base exposed over MCP. Use whenever reading, listing, searching, editing, or linting any `.md` or `.mdx` file in the project, and before any `mcp__open-knowledge__*` tool call (`exec`, `search`, `write`, `edit`, `lint`, and the rest). Installed by `ok init`, so its presence means this is an OpenKnowledge project and it governs every markdown file here. Covers the read/write tool surface, grounding and linking rules, folder/template conventions, the live browser preview, and the rule that OK's MCP tools — never native file tools — handle in-scope markdown.

Use when the user wants to create, author, write, or design a new Agent Skill (a SKILL.md) — for OpenKnowledge or for their editors — including requests like 'help me write a skill', 'make a skill that…', 'turn this workflow into a skill', or improving an existing skill's triggering and discipline. Also use when capturing reusable agent guidance that should live as an installable skill rather than a one-off prompt. Covers choosing scope (project vs global), authoring inside a plugin or skills-distribution repo (write in the repo's own layout, never install), the SKILL.md frontmatter contract, progressive-disclosure structure, evaluating the skill, and installing it into the user's editors.

Evidence-first coding workflow. Use this skill for ANY task that writes, modifies, fixes, refactors, or reviews code — bug fixes, new features, refactors, config changes, even one-line edits. Verifies before presenting, attacks its own output with adversarial multi-model review, and tracks every check in a SQL ledger. The skill's own task-sizing scales effort down for trivial changes, so it is safe to apply broadly. Trigger whenever the user asks for code changes, mentions a bug, feature, refactor, or asks you to implement, fix, or improve anything in a codebase.

Operate the two Fly.io apps for Open Executive (openexec-api-dev, openexec-ui-dev). Use this skill whenever the user asks to tail logs, SSH for SQLite inspection, list/set/unset secrets, check deploy status, restart machines, deploy, or rotate AUTH_SECRET. Covers the exact commands, SQLite paths, and known gotchas for this project. Tiered safety: read-only runs freely; destructive needs explicit user confirmation.

Interact with the Open Executive FastAPI backend via curl. Use this skill when the user asks to hit /chat, /today, /people, /scheduled_actions, /architecture/*, /health/*, /fixtures/*, /audit/*, or any HTTP endpoint on openexec-api-dev.fly.dev (dev) or localhost:8000 (local). Authenticates via $BACKEND_SHARED_SECRET in the x-api-key header. Tiered safety: GET runs freely, mutating POSTs need explicit confirmation.

Operate the Telegram Search CLI through an authorized local profile. Use when an agent needs to discover Telegram chats, read bounded remote messages, explicitly synchronize history with user-approved Telegram Takeout, query/search/contextualize persisted messages, compute statistics, export deterministic JSONL, or run a real CLI end-to-end canary. Never send or modify Telegram messages.

Reverse-engineer design systems, tokens, and components from live products or screenshots

Multi-AI requirements scoping using available external providers (Double Diamond Define phase)

Multi-AI validation, scoring, and review using available external providers (Double Diamond Deliver phase)

Multi-AI implementation using available external providers (Double Diamond Develop phase)

Multi-AI research using available external providers (Double Diamond Discover phase)

Decompose and execute large changes, migrations, or multi-issue fixes in parallel with quality gates

NLSpec authoring — use when you need a structured specification from multi-AI research and consensus

System architecture and API design with multi-AI consensus — use for design reviews and new subsystems

Audit codebases for quality, consistency, and broken patterns — use for pre-release or tech debt review

OpenClaw instance administration — manage hosts across macOS, Ubuntu/Debian, Docker, OCI, and Proxmox

Expert multi-AI code review with inline PR comments — use for thorough quality and security analysis

Extract patterns and anatomy from URLs — use to reverse-engineer content strategies from live pages

Auto-detect work context (Dev vs Knowledge) — use to tailor workflows based on current task type

GitHub Copilot CLI as optional zero-cost provider via copilot -p programmatic mode

Project remaining workflow cost from per-phase averages — warns on budget ceiling overruns

Trace codepaths in diffs, map against tests, auto-generate missing coverage — use before shipping PRs

Structured multi-provider AI debates between Claude and available advisors — use for critical decisions

Debug issues methodically — use when stuck on errors, test failures, or unexpected behavior

Present options with trade-offs for informed decision-making — use when choosing between approaches

Generate slide deck presentations from briefs — use when you need slides, pitch decks, or visual summaries

Thorough research across multiple sources — use for complex topics needing broad synthesis

Persist design documents with branch tracking, revision chains, and cross-session discovery

Audit whether a multi-agent setup earns its coordination cost — use before adding an agent, or when a workflow feels slow or agents agree without adding signal

Principles for writing skills that behave the same way every run — use when adding, editing, or reviewing a skill in this plugin

Convert markdown to DOCX, PPTX, XLSX, PDF office documents — use when you need exportable deliverables

Post-ship doc sync across project markdown. Use when: sync docs, update docs, document changes, release notes.

Environment diagnostics — check providers, auth, config, hooks, scheduler, and more

Run a full build-and-ship pipeline from a spec — use for hands-off project generation

Wrap up a branch — run tests, create PR, merge or discard — use when implementation is done

Move incoming issues and pull requests through triage states until each is actionable or closed — use when the queue has piled up or a report arrives unsorted

Use when starting a complex or ambiguous task that risks scope drift

Track project blockers, bugs, and gaps across sessions — use when issues pile up or need triage

Run tasks in a loop until goals are met — use for iterative refinement, polling, or convergence

Switch to Knowledge Work mode for research and writing — use when task is non-code focused

Craft better prompts using proven optimization techniques — use when your prompt needs refinement

Use when choosing native or multi-LLM handling for init, review, or security requests

Decompose large tasks across parallel agents — use for migrations, multi-file refactors, or batch work

Write an AI-optimized PRD using multi-AI orchestration — use when scoping a new feature or product

Quick execution for ad-hoc tasks without full workflow overhead — use for small, self-contained requests

Pick up where you left off from a previous session — use after context resets, compaction, or new conversations

Use when a reviewer, CI bot, or another AI leaves feedback to address

Roll back to a previous checkpoint via git — use when a change went wrong and you need to revert

OWASP compliance, vulnerability scanning, and adversarial red team testing — use for security reviews

URL validation and content sanitization for untrusted sources — use when handling external input safely

Package and finalize completed work for delivery — use when a feature is done and ready to ship

Use when a PR or feature needs both specification and code-quality review

Show where you are in the workflow and what to do next — use for progress checks and orientation

Manage tasks with Claude Code native tools — use to track TODOs, delegate work, and monitor progress

Build features with tests-before-code rigor — use for new features needing test coverage

Brainstorm creatively with pattern spotting and paradox hunting — use for ideation and exploration

Design UI/UX systems with style guides, palettes, typography, and component specs for new interfaces

Use when about to declare work complete, fixed, passing, or done

Process screenshot-based UI/UX feedback to fix visual issues — use when users share screenshots of bugs

Break a plan or spec into vertical slices that each declare what blocks them — use when work is agreed but not yet cut into fileable pieces

Create zero-context implementation plans with bite-sized tasks — use for multi-step feature planning

Configure Claude Octopus — redirects to /octo:setup interactive wizard

Turn an investment agent into a supply-chain bottleneck hunter. Use this skill for source-backed investment research, live market/theme scans, AI/semi/technology value-chain mapping, A-share/HK/US stock screening, thesis stress tests, and Serenity-inspired research conversations. Trigger on requests like "用 Serenity 的方式看", "深度调研", "产业链/供应链/卡点/瓶颈", "A股 AI 半导体哪个最值得研究", "find unknown bottlenecks", "rank candidates", or "challenge this thesis". Outputs plain-language reasoning, ranked research priorities, evidence chains, risks, and next verification steps. Research support only; no trade execution.

Plug-in vision for text-only models. Hard rule: when a file path or URL with an image extension (.png, .jpg, .jpeg, .webp, .gif, .heic, .heif) appears anywhere in the conversation (typed by the user, injected as a `[Image: source: <path>]` line, or inside a tag) and you cannot see that image's content, run this skill on it before any other approach: no self-built OCR, no PIL, no tesseract. Also triggers on pasted-image placeholders such as `[Image #1]` and `[Unsupported Image]`. If you can actually see the image, do not use this skill. When unsure, run `modlens guard` before the first read of a session: a deny verdict means the active model has native vision and must read the image itself. Runs the modlens CLI to convert the image into structured JSON evidence: every word transcribed, layout regions, semantics, visual clues. Also use when the user asks how to install, configure, or switch modlens providers (Gemini API key, OpenAI-compatible endpoints, Claude API or Claude Code CLI).

Use when building, extending, or debugging AI agents with cascadeflow (agent runtime intelligence layer) — installing `cascadeflow` (Python) or `@cascadeflow/core`/`@cascadeflow/langchain` (TypeScript); using `CascadeAgent`, `ModelConfig`, harness APIs (`cascadeflow.init`, `cascadeflow.run`, `@agent` from `cascadeflow.harness`, `simulate`), `withCascade`/`CascadeFlow`; picking drafter+verifier pairs; per-step budget/compliance/KPI enforcement; quality validation; complexity pre-routing; tool execution and multi-turn agent loops; presets; decision traces; or wiring cascadeflow into LangChain, OpenAI Agents, CrewAI, PydanticAI, Google ADK, n8n, or Vercel AI SDK. Also when a user mentions "cascade", "drafter/verifier", "runtime intelligence", "in-process harness", "cost-optimized agent", "agent loop with cost control", is in the lemony-ai/cascadeflow repo, or found a bug in cascadeflow/integrations needing an upstream fix/PR.

|

Agent Context Isolation

Agent Orchestration Rules

Agentic Workflow Pattern

Guide for integrating Agentica SDK with Claude Code CLI proxy

Reference guide for Agentica multi-agent infrastructure APIs

Write reliable prompts for Agentica/REPL agents that avoid LLM instruction ambiguity

Build Python agents with Agentica SDK - @agentic decorator, spawn(), persistence, MCP integration

Agentica server + Claude proxy setup - architecture, startup sequence, debugging

Spawn Agentica multi-agent patterns

Async REPL Protocol

Semantic search across codebase using LEANN vector index

Router-First Architecture

Search Tool Hierarchy

Search Tool Hierarchy

AST-based code search and refactoring via ast-grep MCP

Background Agent Pings

Analyze Claude Code sessions via Braintrust

Braintrust tracing for Claude Code - hook architecture, sub-agent correlation, debugging

Workflow orchestrator that chains existing skills for feature development

Claude Code CLI commands, flags, headless mode, and automation patterns

Create git commits with user approval and no Claude attribution

A complete skill for E2E testing

Completion Check: Verify Infrastructure Is Wired

Transform session learnings into permanent capabilities (skills, rules, agents). Use when asked to "improve setup", "learn from sessions", "compound learnings", or "what patterns should become skills".

Create or update continuity ledger for state preservation across clears

Create handoff document for transferring work to another session

Find unused functions and dead code in the codebase

Systematic hook debugging workflow. Use when hooks aren't firing, producing wrong output, or behaving unexpectedly.

Debug issues by investigating logs, database state, and git history

Generate comprehensive PR descriptions following repository templates

Deep interview process to transform vague ideas into detailed specs. Works for technical and non-technical users.

Environment Triage