Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Unified PDCA cycle management — plan, design, do, analyze, iterate, report. PDCA runs per-feature (9-phase: pm→plan→design→do→check→act→qa→report→archive); for multi-feature scope/budget grouping use /sprint (v2.1.13, 8-phase container that may host PDCA cycles inside). Triggers: pdca, plan, design, analyze, report, status, next, iterate
Autonomous build-phase orchestrator with 8-stage per-slice pipeline (decompose, slice readiness review, implement via TDD pairs, three-stage code review, address feedback, mutation testing, push/CI, merge/flag). Manages slice queue (walking skeleton first, dependency ordering), dispatches TDD pairs with capability detection, enforces 5 binary quality gates with 3-cycle rework budgets and human escalation, maintains audit trail, and respects strict controller boundaries (never writes code, delegates all creative work). Supports three autonomy levels (conservative/standard/full). Use when running factory mode, orchestrating build phases, managing slice queues, or coordinating TDD pairs with review gates. Triggers on: "run the pipeline", "factory mode", "build phase", "slice queue", "quality gates", "autonomous build". NOT for: individual TDD (use tdd), standalone code reviews (use code-review), CI-only workflows (use ci-integration).
Orchestrated, idea-to-plan PM workflow. Establishes context once, then composes the PM generative ladder in order — problem → discovery → users → riskiest-assumption test → prioritization → spec → metrics — gating each step, enforcing evidence before the spec, and handing the result to product-review. Trigger when the user wants to go from a raw idea to a build-ready plan ("take me from problem to plan", "I have an app idea, where do I start", "turn this idea into a product plan", "plan a new product end to end") — but NOT when they ask for one specific PM artifact (let that skill trigger directly), and NOT for the strategy/launch arc (vision, competitive analysis, roadmap, GTM — offer those next).
Runs autonomous red-green-refactor loops to fix failing tests, reach coverage targets, and satisfy quality gates. Use when tests need to pass, coverage thresholds must be met, quality gates require compliance, or flaky tests need stabilization.
Implement code against an existing Quint specification. Uses Research → Plan → Implement workflow (ACE-FCA style) grounded by the spec as the source of truth. Use when the user wants to refactor code, add a new feature, or close a gap between implementation and spec — with the Quint spec as the formal constraint that all changes must satisfy.
Operate any desktop software using "screenshots + coordinate grid": capture a screenshot with a grid, let any multimodal LLM read out the target's coordinate numbers, then call click/type/keypress APIs to simulate how humans visually operate software. It can also record an operation once and save it as a reusable scenario template for direct reuse next time. Does not rely on the target software providing an API or CLI. Use it when: - Automating Windows desktop software (WeChat, heavily risk-controlled web pages such as Xiaohongshu/Douyin, phone emulators, games, any Windows program), especially software or elements that Playwright/CDP/CLI/application APIs cannot cover (such as hidden buttons, game screens, CAPTCHAs) - The user says things like "help me operate/click/type XX", "automate XX", "simulate clicks", "desktop automation", "RPA", "record operations", "save as a template", "run automatically next time" - Operating desktop elements or cross-process, cross-window elements (desktop icons, taskbar, tray, Start menu, file selection dialogs, cross-window drag and drop) As long as the target is a Windows desktop program that can be seen on screen and requires mouse and keyboard operation, this skill should be tried first even if it has no API, no dedicated skill, or an unclear automation path.
Build, configure, and test Agentforce agents on Salesforce. Covers agent setup, topics, actions (Flow, Apex, PromptTemplate, External Service), Agent Scripts for deterministic FSM-based agents, PromptTemplate authoring, GenAI Models API, metadata structure, and agent testing via the Testing Center and CLI. Activate on Agentforce agents, .agent files, .agent-meta.xml, .agentTopic-meta.xml, .agentAction-meta.xml, GenAiFunction, GenAiPlugin, PromptTemplate, Agent Script, or agent testing workflows.
Keep a reviewable decision trail for long-running or unattended work: a TSV log with one row per decision (what, why, evidence, result). Local by default; commit it when a reviewer needs the trail to trust the result. Use for /show-me-your-work, autonomous or multi-phase runs, or work a human reviews after stepping away.
Manage duoduo's idle auto-compact for channel sessions — enable/disable per conversation, read the measured compaction stats, and retune thresholds with the break-even formula when the owner changes model/backend or asks about idle-session cost. Trigger when the owner mentions 自动压缩/smart compaction/压缩策略, asks why a session was compacted (a compact notice appeared), switches a session to a metered or short-cache-TTL backend, or asks to protect a conversation from compaction. Also use proactively right after the FIRST auto-compaction of a session you manage — that event carries the calibration measurement.
Eight-stage PDCA workflow for systematically solving bugs, refactors, and feature-development tasks: clarify → analyze → explore solutions → review → plan → execute → verify → retrospect. Manual mode (default) pauses for user confirmation at each stage exit; auto mode runs end-to-end; ai-proxy overlay (thin freeze then occupy) is available independently. Triggers — 「明确问题」「分析问题」「探索方案」「审查方案」「制定计划」「执行计划」「检查验证」「复盘改进」(alias「回顾总结」)；「继续分析」「深入分析」「修改方案」「完善方案」「优化方案」「更新计划」「修订计划」「修改计划」；「自动模式」「自动分析」「自动解决」；「ai-proxy 模式」「AI 代理模式」「切换 ai-proxy」 / clarify problem, analyze problem, explore solutions, review solution, make plan, execute plan, verify, retrospective, auto mode, ai-proxy mode, switch to ai-proxy.
Helps coding agents use vit to discover, follow, skim, and ship software capabilities (caps) over ATProto. Activates when the user mentions vit, beacons, caps, shipping, skimming, following, vetting, or social coding. Provides command reference for: vit init, vit skim, vit ship, vit follow, vit following, vit learn, vit remix, vit doctor, vit config, vit beacon, vit login.
Analyze mini-program project source code (including minified/obfuscated code), identify core business steps, extract network interfaces and JSAPI calls, generate a skill subpackage compliant with the wx.modelContext specification (with atomic interfaces + atomic components), and complete the app.json / project.config.json configuration integration. Trigger in these scenarios: turning mini-program page capabilities into mini-program AI atomic interfaces, generating skills/ subpackage code, deriving MCP tools from a source project, code generation for mini-program AI development mode. Handles static generation only; after generation completes, hand off to wxa-skills-validate for validation.
Prepares Go applications to work effectively with AI coding agents. Use when setting up a new Go project or retrofitting an existing one to ensure: structured JSON logging (slog/Zap/ZeroLog/Logrus), machine-readable command output, thorough golangci-lint configuration, non-interactive CLI design with --yes flags, structured error handling with meaningful exit codes, proper context.Context propagation, graceful shutdown, health check endpoints, and a standardized Makefile. Triggers when a user asks to make their Go app "agent-ready," "AI-friendly," wants to improve agent tooling/observability in a Go project, or needs to audit an existing Go project against agent-readiness best practices.
Stop your AI from making things up. Use when your AI hallucinates, fabricates facts, is not grounded in real data, does not cite sources, makes unsupported claims, or you need to verify AI responses against source material. Also use when your LLM makes up facts, responses are disconnected from the input, or outputs are not grounded in source documents. Covers citation enforcement, faithfulness verification, grounding via retrieval, confidence thresholds, and evaluation of anti-hallucination quality. Also used for AI makes up citations, LLM fabricates data, ground AI in source documents, RAG but AI still hallucinates, force AI to cite sources, factual accuracy for AI, prevent AI from inventing facts, AI confident but wrong, LLM confabulation, hallucination detection, verify AI claims against documents.
AI tooling for Bitrix24 development: official docs MCP b24-dev-mcp (bitrix-search, bitrix-method-details) and Alaio Vibecode (keys, /v1/me, vibecodeconnector on-premise). Use when an agent writes REST calls, sets up MCP, or builds a Vibecode app.
Orchestration spine and "what do I do next?" router for the BMAD Planning & Orchestrator plugin. Detects current planning state by scanning the output folder for which artifacts exist (project-context.md, decision-log.md, product-brief, prd/tech-spec, architecture, ux, epics, stories), infers the current phase and track, then recommends the next BMAD skill to run — skipping optional phases (Analysis is always optional; UX only when there is a UI). Use when the user asks "what's next", "what do I do next", "where am I", "what's my status", "continue", "resume planning", "which skill should I run", "what BMAD step is next", or types "/bmad-help" / "/status". Auto-invoke at the start of any BMAD planning session to orient. This skill ROUTES only — it never writes planning documents itself.
Use when creating new opencode skills, modifying existing skills, or measuring skill performance. Triggers include requests to create a skill from scratch, edit or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.
Three-agent adversarial debate protocol for strategic decisions. Two advocates steelman opposing positions while a moderator identifies risks and synthesizes a binding consensus. Use when choosing between alternatives, evaluating tradeoffs, or making high-stakes decisions.
Build DSPy 3.2.x programs through spec, program, metric and baseline; extend to optimization and export when requested and justified by task budget. Orchestrates the other four DSPy skills (dspy-fundamentals, dspy-evaluation-harness, dspy-gepa-optimizer, dspy-rlm-module) in the correct order. Use for greenfield DSPy builds; prototypes may stop at a validated baseline.
Write idiomatic DSPy 3.2.x programs — typed Signatures, dspy.Module subclasses, Predict/ChainOfThought/ReAct/ProgramOfThought, and save/load. Use this when starting any new DSPy project or when fixing non-idiomatic DSPy code (hard-coded prompts, ad-hoc string templates, untyped outputs, non-serializable classes).
Drive the Duvo public API from the terminal via the `duvo` CLI (`@duvoai/cli`). Use when the user wants to script Duvo — managing automations, Policy, agents, runs, cases, queues, files, skills, connections, Clarity processes, Pulse dashboards, or hitting an arbitrary endpoint via `duvo api` — instead of clicking through the Duvo web UI or hand-crafting `curl` calls.
CodeStable is an AI coding workflow for serious software engineering that keeps humans in the loop. Rather than centering on Agent orchestration, it organizes requirements, architecture, features, issues, and historical decisions to make development driven by Codex and Claude controllable, traceable, and continuously evolvable.
CodeStable is an AI coding workflow for serious software engineering that keeps humans in the loop. Rather than centering on Agent orchestration, it organizes requirements, architecture, features, issues, and historical decisions to make development driven by Codex and Claude controllable, traceable, and continuously evolvable.
Automate Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify, etc.) using agent-browser via Chrome DevTools Protocol. Use when the user needs to interact with an Electron app, automate a desktop app, connect to a running app, control a native app, or test an Electron application. Triggers include "automate Slack app", "control VS Code", "interact with Discord app", "test this Electron app", "connect to desktop app", or any task requiring automation of a native Electron application.
Bidirectional sync between BMAD planning artifacts and GitHub Projects v2. Use this skill whenever the user says "sync to github", "push stories", "pull status from github", "set up github project", "github sync", "onboard github project", "create github issues from stories", "update stories from github", "sync status", or wants to manage BMAD artifacts in GitHub Projects. Also trigger when user mentions "gh project", "github board", "sprint board", "project roadmap", "relationship field", "blocked by", or "parent issue" in the context of BMAD artifacts. Trigger on implicit cues like "I want to track stories in github", "let's set up a project board", "what's out of sync", "push the sprint to github".
Generates implementation plans with milestones and tasks from business and technical requirements. Supports AI, human, and hybrid audiences with appropriate detail levels. Embeds best practices including task sizing (30m-2.5h), style anchors, TDD requirements, and quality constraints. Outputs milestones.yaml and milestone-m*.tasks.yaml files ready for development.
Use when setting up or configuring Laravel Boost for AI-assisted development — package installation, MCP server configuration, guideline customization, skill authoring, documentation API integration. Trigger conditions: install Laravel Boost, configure MCP for IDE, create custom AI guidelines, write project-specific skills, verify MCP tool connectivity, update Boost after dependency changes, extend Boost for custom agents.
Extracts lessons learned from conversations and persists them to AI assistant config files (CLAUDE.md, GEMINI.md, AGENTS.md, Cursor, Copilot, Windsurf, Continue) or creates new skills from discovered workflows. Use this skill whenever: a command failed then succeeded (especially with a flag or env change), debugging revealed a workaround, an assumption proved wrong, undocumented behavior was discovered, a multi-step pattern was invented on the fly, or the user says "/learn", "remember this", "save this", "add this to CLAUDE.md", "don't forget this", or "make a skill for this". Also invoke proactively after any complex troubleshooting session where notable lessons emerged, even if the user didn't explicitly ask.
Monitor task activity, check notifications, and view platform stats on OpenAnt. Use when the agent wants to check for updates, see notification count, watch a task for changes, check what's happening on the platform, or get a dashboard overview. Covers "check notifications", "any updates?", "platform stats", "what's new", "status update", "watch task". For personal task history and listing, use the my-tasks skill instead.
Eight-stage PDCA problem-solving workflow that persists analysis, proposal, design review, plan, execution, and verification into OpenSpec artifacts (openspec/changes/<name>/, archived into openspec/specs/) instead of leaving them only in chat context. Use for feature work, bug fixes, refactors, and complex engineering tasks that need long-term behavioral-contract traceability, team review, or auditability. Do NOT use for a quick one-off edit with no traceability need — use solve-workflow instead. Triggers：「opsx解决」「OpenSpec解决」「规范化解决」「创建OpenSpec变更」「创建opsx变更」「用OpenSpec分析」「用OpenSpec修复」「opsx自动解决」「OpenSpec自动解决」「opsx-solve」「opsx-solve-workflow」；「ai-proxy 模式」「AI 代理模式」「切换 ai-proxy」 / opsx solve, OpenSpec solve workflow, create an OpenSpec change, ai-proxy mode, switch to ai-proxy.
Plan a round of user interviews conducted by PostHog''s AI voice agent (a "robo interviewer") — the automated voice-agent interview product. Captures a UserInterviewTopic (who to target, what to ask, framing context, question list) and calls user-interview-topics-create. ONLY trigger when the user clearly wants an AI voice agent to actually run the interview calls (e.g. "set up robo user interviews", "have the voice agent interview these users"). Do NOT trigger for ordinary user research that does not involve the voice agent — finding or shortlisting users to talk to ("who''d be a good fit to interview about Y"), planning questions for a human-run interview, or analysing feedback are audience discovery, handled with normal data queries, not this skill. Also do NOT trigger for uploading a recorded interview audio file or browsing topics with user-interview-topics-list. When intent is ambiguous, first confirm what kind of research it is and whether they want an AI voice agent to conduct it (see Step 0).
Audit and scaffold `.github/workflows/copilot-setup-steps.yml` for GitHub Copilot cloud agent environments. Use for missing or broken setup workflows, runner/firewall/dependency/LFS/proxy issues, or Copilot setup repair. Do NOT use for generic GitHub Actions.
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.