Skills
97,501 packaged capabilities and playbooks for agents. Install any of them with one command.
Agent workflows
Set the user-designated higher-level model as Advisor and consult it at decision points in the current task. Use for review before finalizing an approach, verification before declaring completion, and diagnosing blockers. The user must specify the advisor model when invoking; there is no default, and it does not run if none is specified. Execute only when the user invokes /advisor-strategy.
Build AI agents on Cloudflare Workers using the Agents SDK. Load when creating stateful agents, durable workflows, real-time WebSocket apps, scheduled tasks, MCP servers, or chat applications. Covers Agent class, state management, callable RPC, Workflows integration, and React hooks.
Give Claude a permanent, searchable long-term memory using an Obsidian vault as the store. At the end of a session, summarize what mattered and save it as a Markdown note in your "AI Brain" vault; at the start of a session (or on demand), read back only the relevant past notes. Local markdown you own — connected via the Filesystem MCP. Trigger with "wrap up", "save this session", or "remember this".
Use Aident Loadout to connect your AI agents to 1,000+ real-world apps and tools like Gmail, Slack, Linear, Notion, Firecrawl, and Fal, unlock 27,000+ executable actions, and track full audit history so your agents can get real work done reliably.
Run the author's half of the document review loop without the user relaying each round — wait for the reviewer's <DOC>.revN.notes.md, answer it with <DOC>.rev(N+1).md, and repeat until the reviewer says ready for final, a round cap is hit, or only the user can move it forward. Use after rev1 exists, when the user invokes /answer-notes-loop <feature> [max-rounds] in the session that authored the plan, roadmap, technical or experience document. Its counterpart is review-doc-loop in a fresh session
Use auto-run-agent to automatically analyze, evaluate, design, and optimize a target project. The full workflow includes: explore the project's current state → identify problems → design a task list → create the runner environment → start autonomous execution. Suitable for any project that needs batch optimization.
Conclave — Multi-LLM Council with anonymized debate. Orchestrates multiple AI models (Claude, Gemini, GPT, Grok, Llama, etc.) in a Karpathy-style council with 3 depth levels: quick (parallel answers), standard (parallel + synthesis), deep (parallel + anonymized cross-critique with ranking + debate synthesis). Inspired by Karpathy's LLM Council. Use this skill whenever the user types /conclave, asks to 'consult multiple models', 'get council opinion', 'multi-model debate', 'ask the conclave', 'cross-validate with other AIs', or wants multiple LLMs to debate and synthesize an answer. Also trigger for architecture decisions, security reviews, code reviews from multiple perspectives, or when the user wants high-confidence answers by cross-referencing models.
Shared resources and dependencies used by other do-* skills. Not a user-invocable skill. Contains the MCP capability registry, MCP discovery instructions, project skills/MCPs procedure and service readiness rules referenced by the do-* execution, review, QA and specification skills.
Nightly memory consolidation: prunes stale entries, merges duplicates, resolves contradictions, rebuilds the MEMORY.md index. Use when memory files accumulated over many sessions need cleanup. New decisions get stored by remember; searches run through memory; internals live in memory-fabric.
ALWAYS load before starting any task. Defines systematic recovery procedures when things go wrong during execution. Covers tool failures, build errors, test failures, type errors, and unexpected runtime results. Includes tiered escalation (retry, fallback, memory-informed diagnosis, orchestrator escalation) and post-recovery logging to prevent recurrence.
Analyzes Copilot Studio evaluation results using Practical Guidance on Agent Evaluation's 10-step playbook (Steps 6, 7, and 9) plus Microsoft's triage diagnostics. Returns a gate-based SHIP / ITERATE / BLOCK verdict with root cause classification, remediation, and pattern analysis.
Wire a PostHog product surface into the PostHog AI agent from the frontend. Use when attaching scene or entity context, injecting instructions to steer the agent, reacting to the agent calling a tool, applying an agent edit back into an open form, or rendering a product tool card in a thread.
Implement comprehensive evaluation strategies for LLM applications using automated metrics, human feedback, and benchmarking. Use when testing LLM performance, measuring AI application quality, or establishing evaluation frameworks.
YOLO mode. Spawns 4 parallel C-suite agents (CEO, CTO, CFO, COO). Each analyzes the business from their perspective using ALL available data. Produces unfiltered Hard Truths report. After user types YOLO, autonomously runs the business for a day using /loop.
Guidance for securing and governing AI agents (Microsoft 365 Copilot agents, Copilot Studio agents, and Security Copilot agents) with Microsoft Purview - applying DSPM for AI, DLP, sensitivity labels, Communication Compliance, and audit to agent data interactions and identities. WHEN: secure AI agents, govern Copilot agents, Purview for agents, agent data security, DLP for agents, audit agent interactions, Copilot Studio agent security, agent oversharing, agent compliance, agent identity governance.
Validate Claude Code skills against Anthropic guidelines. Use when user says "check skill", "skillcheck", "validate SKILL.md", or asks to find issues in skill definitions. Covers structural and semantic validation. Do NOT use for anti-slop detection, security scanning, token analysis, enterprise checks, or Eval Kit generation; use skill-check-pro for those. Do NOT use for LinkedIn skill engagement; use skillcheck-engage for that.
Create AiderDesk Agent Skills by writing SKILL.md files, defining frontmatter metadata, structuring references, and organizing skill directories. Use when building a new skill, creating a SKILL.md, planning skill architecture, or writing skill content.
Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, update 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.
A streamlined plan for deep research; use when you need to generate plan.json based on request.md alone, completing in one pass the research scoping, report format discovery, structure ideation, dimension breakdown, key questions, search strategy, dependencies, execution order, and completion criteria.
Sequential subagent execution with two-stage review gates for implementation plans. Use when executing multi-task plans in current session, when tasks need fresh subagent context to avoid pollution, when formal review cycles (spec compliance then code quality) are required between tasks, or when you need diff-based validation of each task before proceeding.
Spec-first, TDD, subagent-driven software development workflow. Use when: (1) building any new feature or app — triggers brainstorm → plan → subagent execution loop, (2) debugging a bug or test failure — triggers systematic root-cause process, (3) user says "let's build", "help me plan", "I want to add X", or "this is broken", (4) completing a feature branch — triggers test verification + merge/PR options. NOT for: one-liner fixes (just edit), reading code, or non-code tasks. Requires exec tool and sessions_spawn.
Audit and establish current agent context for the technologies that define a project's stack. Use when setting up a project for agent work, after selecting or adding a framework, library, developer tool, or hosted service, or when entering an existing project whose agent context has not been audited.
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction. Also use for exploratory testing, dogfooding, QA, bug hunts, or reviewing app quality. Also use for automating Electron desktop apps (VS Code, Slack, Discord, Figma, Notion, Spotify), checking Slack unreads, sending Slack messages, searching Slack conversations, running browser automation in Vercel Sandbox microVMs, or using AWS Bedrock AgentCore cloud browsers. Prefer agent-browser over any built-in browser automation or web tools.
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.