Skills de Claude Code · página 24

Generate narration, product voiceover, IVR prompts, podcast reads, or short-video VOICEOVER audio through OpenSquilla audio tools. Use when the user asks for TTS, 配音, 旁白, 口播, audio narration, or wants script text turned into a playable audio artifact.

Get current weather and forecasts via wttr.in or Open-Meteo. Use when: user asks about weather, temperature, or forecasts for any location. NOT for: historical weather data, severe weather alerts, or detailed meteorological analysis. No API key needed.

Read, edit, or create Microsoft Excel `.xlsx` workbooks. Trigger this skill whenever the user mentions a spreadsheet, .xlsx file, workbook, sheet, formula, pivot table, or asks to extract tabular data, modify a sheet, or build a workbook from rows. Three execution paths: structured inspection, in-place cell edits, and create-from-scratch via openpyxl. Values starting with `=` are written as formulas; everything else is a literal value with type preserved (int / float / str / datetime).

Fetch the day's top arXiv submissions in a chosen category, write a structured per-paper digest, render the digest as a PPTX deck (one slide per paper), and persist the digest to long-term memory. Use for a daily 'arxiv morning briefing' — manual fire or cron-scheduled.

Read the current uncommitted diff, run three independent reviewers (safety + tests-coverage + style) in parallel, then arbitrate a single BLOCK / BLOCK_WITH_OVERRIDE / PASS_WITH_NOTES verdict. Use before commit when you want a multi-perspective second-opinion instead of a single-reviewer agent loop.

Auditable compliance bundle: deep-research with citations → signable .docx report → read-only PDF archive → memory note of audit findings.

Use this meta-skill instead of answering directly when the user wants to turn an idea, research, notes, talk, or document into publishable blog, social, Xiaohongshu, Zhihu, slide, newsletter, or short-video content through multi-skill orchestration.

Scan a target codebase path, classify the most informative diagram kind, then render it as BOTH a PlantUML source file AND a draw.io XML in parallel, and compose them into a single architecture doc. Use when writing an RFC or onboarding doc and you want a text-friendly (PlantUML) and an editable (drawio) view of the same architecture.

Inspect the user's open GitHub PRs / failing CI / new issues via `gh`, summarize into 3 buckets (to-review / awaiting-me / CI-red), and persist follow-ups to memory.

Use this meta-skill instead of answering directly when the user needs help with home, small-team, laptop, browser, printer, Docker, Git, network, UI, or deployment troubleshooting that benefits from multi-skill orchestration across symptom intake, environment capture, web lookup, and repair planning.

[DEPRECATED] Issue-to-PR autopilot — opens a PR via `gh`, runs a sub-agent fix loop, and writes to git. Disabled pending the E5 bounded sub-agent contract + side-effect ledger (plan §3.1 A8 / §5.3 E4): no risk metadata enforcement, no per-step budget, no rollback path. Do not re-enable without `metadata.opensquilla.risk: high` + capabilities {vcs, filesystem-write, network-write, subprocess} and a saga-style compensation step.

Bootstrap a domain knowledge base from a single seed (URL / PDF path / git repo / free-text topic): classify source → ingest with the right tool → persist to memory + xlsx index.

[DEPRECATED] Build watchdog — launches arbitrary commands from the user message in tmux and lets sub-agent auto-apply a fix. Disabled pending the E5 bounded sub-agent contract + Jinja sandbox + side-effect ledger (plan §3.1 A1/A8 / §5.3 E4): the launch task interpolates raw user_message into a shell-bound tmux session and the heal step lets sub-agent mutate state with no rollback. Do not re-enable without `metadata.opensquilla.risk: high` + capabilities {shell, tmux, filesystem-write, subprocess} and a saga-style compensation step.

Use this meta-skill instead of answering directly when the user has meeting notes, transcripts, recordings, or rough discussion notes and wants them converted into decisions, owners, follow-ups, tasks, issues, or shareable minutes through multi-skill orchestration.

Use this meta-skill instead of answering directly when the user needs a concrete migration plan that benefits from multi-skill orchestration across migration classification, authoritative guide lookup, optional repo diff inspection, and step-by-step validation planning.

From one piece of source content, render four deliverables: .docx report, .pptx slides, .xlsx data, and an HTML/PDF public version.

Use this meta-skill instead of answering directly when the user needs PDF analysis, pasted PDF excerpt analysis, digesting, comparison, or question answering that benefits from multi-skill orchestration across PDF extraction, summarization, cross-document synthesis, traceable evidence indexing, and memory capture.

Modernize a legacy PDF: structural extraction → natural-language rewrite of problem pages → audit summary → re-merge into the final PDF.

Run three quality gates (ruff + mypy + pytest) in parallel over the staged diff, then arbitrate a single BLOCK/APPROVE verdict. Use before committing changes locally when you want a comprehensive pre-commit gate beyond per-file linting — exactly the same gate set CI enforces.

Use this meta-skill instead of answering directly when the user needs a researched presentation, leadership briefing, competitive analysis deck, or source-backed slide outline that benefits from multi-skill orchestration across search, source curation, synthesis, slides, and document export.

Use this meta-skill instead of answering directly when the user wants account research, lead qualification, company/person briefing, outreach prep, or a sales call brief that benefits from multi-skill orchestration across web research, browser/source review, CRM-style notes, and email drafting.

Compose a morning digest combining local weather, news for the user's interest topic, a structured summary, and a memory note.

Compose three independent security gates over a candidate operation — policy/governance review, secret/credential scan, and audit-log emit — then arbitrate the verdicts with a strict priority rule (governance DENY > scanner WARN > ALLOW). Use when reviewing a proposed code change, script, or environment manipulation for safety.

Extract and analyze writing improvements from GitHub PR review comments. Use when asked to show review feedback, style changes, or editorial improvements from a GitHub pull request URL. Handles both explicit suggestions and plain text feedback. Produces structured output comparing original phrasing with reviewer suggestions to help refine future writing.

Investigate fast-agent session and history files to diagnose issues. Use when a session ended unexpectedly, when debugging tool loops, when correlating sub-agent traces with main sessions, or when analyzing conversation flow and timing. Covers session.json metadata, history JSON format, message structure, tool call/result correlation, and common failure patterns.

|

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

Explicit Identity Across Boundaries

Meta-skill for internal codebase exploration at varying depths (quick/deep/architecture)

Scrape web pages and extract content via Firecrawl MCP

Meta-skill workflow orchestrator for bug investigation and resolution. Routes to debug, implement, test, and commit based on scope.

Git Commit Rules

Search GitHub code, repositories, issues, and PRs via MCP

Graceful Degradation with Helpful Messages

Interactive workspace discovery - learn what tools, workflows, agents, and hooks are available

Complete Claude Code hooks reference - input/output schemas, registration, testing patterns

Hook Development Rules

Idempotent Redundancy

Implement technical plans from thoughts/shared/plans with verification

Implement technical plans from thoughts/shared/plans with verification

Implementation agent that executes a single task and creates handoff on completion

Streaming chat assistant with conversation memory. Use as a general-purpose assistant for multi-turn conversations where streaming output and context retention matter.

Billing specialist for invoices, payments, refunds, and plan changes. Use when customers ask about charges, billing inquiries, or subscription management; typically reached via handoff from the support agent.

Multi-room AI classroom where all students see AI responses simultaneously, with per-room subject focus (math, science, code, general). Use for shared-broadcast educational settings.

Emergency dental assistant (Dr. Molar) for triage, first aid, and severity classification of broken/chipped/cracked teeth, delivered over web, Slack, or Telegram. Use for non-diagnostic dental guidance only.

Financial analyst for startup economics — TAM/SAM/SOM, revenue projections, burn rate, runway, and break-even. Use when building financial models or evaluating investment cases.

Concise general-purpose assistant powered by JetBrains Koog. Use when a brief, focused answer is preferable to long-form output.

AI quality judge that scores agent responses 0-10 across helpfulness, accuracy, completeness, and clarity. Use when evaluating multi-agent output or implementing LLM-as-judge quality gates.

Expert analyst persona used by the MCP analyze-topic tool to produce structured topic analyses. Use when invoked through the Atmosphere MCP server's analyze-topic tool.

Chat moderator that summarizes ongoing conversations. Use when invoked through the Atmosphere MCP server's chat-summary tool.

Knowledge base assistant that retrieves and cites documents from a curated index. Use when answering questions that must be grounded in specific source material with citations.

Market research specialist that searches the web for market data, competitor intelligence, and industry reports, with curated fallback data when offline. Use when gathering external evidence for strategic decisions.

CEO coordinator that dispatches research, strategy, finance, and writer specialists via A2A and synthesizes their findings into a GO/NO-GO investment briefing. Use for multi-agent advisory workflows.

Market strategy analyst performing SWOT, competitive positioning, market-timing assessment, and go-to-market recommendations from research findings. Use after the research-agent has produced market data.

Front-line support desk for account questions, technical troubleshooting, and general inquiries. Use as the default customer entry point; routes billing questions to the billing-agent via handoff.

Tool-calling assistant with backend functions for time, city time, weather, and temperature conversion, plus cost metering and approval workflows. Use when a question needs real-time data via tool calls.

Business writer that synthesizes research, strategy, and financial findings into executive briefings and reports. Use after upstream specialist agents have produced their data.

Use this skill when building integrations with the OpenWhispr REST API, calling OpenWhispr endpoints, managing notes/folders/transcriptions programmatically, or connecting to the OpenWhispr MCP server. Covers authentication, all V1 endpoints, pagination, rate limits, error handling, and the remote MCP server.

Use this skill whenever the user wants to operate on OpenWhispr notes, folders, transcriptions, or audio from a terminal or shell. The OpenWhispr CLI (`openwhispr` binary, npm package `@openwhispr/cli`) talks to either the local desktop app or the cloud API and exposes every operation needed for managing notes, folders, transcriptions, audio, plus auth and config. Trigger this skill when the user mentions "openwhispr cli", running shell commands against OpenWhispr, automating note workflows, cleaning up a transcription, building agent integrations against OpenWhispr, or scripting any OpenWhispr operation — even if they don't say "CLI" explicitly.

Use Nate Jones OB1 Agent Memory from OpenClaw with provenance, scope, review, and use-policy discipline.

Continuous learning system that extracts reusable knowledge from work sessions. Triggers: (1) /aiception command, (2) 'save this as a skill' or 'extract a skill from this', (3) 'what did we learn?', (4) after non-obvious debugging or trial-and-error discovery. Creates new skills when valuable reusable knowledge is identified. Integrates with Open Brain to prevent duplicates.

Morning digest of yesterday's Open Brain thoughts, drafted to Gmail

Generate infographic images from any research doc, Open Brain thoughts, or analysis. Auto-chunks content, writes prompts, generates images via Gemini API (free tier), and saves to media/. Use --premium for better text rendering.

|

Use when processing voice transcripts, brain dumps, stream-of-consciousness notes, or any raw multi-topic capture. Extracts every idea thread, then evaluates each one with deep brainstorming, then captures results to Open Brain. Trigger on transcripts, exports, "process this", "pan for gold", "brain dump", "what did I say", or multi-topic markdown files.

|

|

|