Agent Workflow Claude Skills
4 121 agent workflow skills from 665 authors. They configure the agents themselves: memory, prompts, context and other skills. Half of them fit into 1 845 tokens or less — that is what one costs your context window when the agent loads it. 778 ship runnable scripts rather than instructions alone. 5 of them cannot work without an MCP server, most often task. We also found 541 copies of these same skills sitting in other people's repositories — counted once here, not 541 times.
4 121 unique 665 authors 2 764 updated this month 506 from vendors
基于RFC驱动的多智能体DAG执行模式，包含质量门、合并队列和工作单元编排。
Head-to-head comparison of coding agents (Claude Code, Aider, Codex, etc.) on custom tasks with pass rate, cost, time, and consistency metrics
Design and optimize AI agent action spaces, tool definitions, and observation formatting for higher completion rates.
Patterns and architectures for autonomous Claude Code loops — from simple sequential pipelines to RFC-driven multi-agent DAG systems.
Orchestrate multi-agent coding tasks via Claude DevFleet — plan projects, dispatch parallel agents in isolated worktrees, monitor progress, and read structured reports.
Audits Claude Code context window consumption across agents, skills, MCP servers, and rules. Identifies bloat, redundant components, and produces prioritized token-savings recommendations.
Patterns for continuous autonomous agent loops with quality gates, evals, and recovery controls.
Instinct-based learning system that observes sessions via hooks, creates atomic instincts with confidence scoring, and evolves them into skills/commands/agents.
Automatically extract reusable patterns from Claude Code sessions and save them as learned skills for future use.
Multi-agent orchestration using dmux (tmux pane manager for AI agents). Patterns for parallel agent workflows across Claude Code, Codex, OpenCode, and other harnesses. Use when running multiple agent sessions in parallel or coordinating multi-agent development workflows.
Operate long-lived agent workloads with observability, security boundaries, and lifecycle management.
Formal evaluation framework for Claude Code sessions implementing eval-driven development (EDD) principles
Pattern for progressively refining context retrieval to solve the subagent context problem
Build MCP servers with Node/TypeScript SDK — tools, resources, prompts, Zod validation, stdio vs Streamable HTTP. Use Context7 or official MCP docs for latest API.
Scan your Claude Code configuration (.claude/ directory) for security vulnerabilities, misconfigurations, and injection risks using AgentShield. Checks CLAUDE.md, settings.json, MCP servers, hooks, and agent definitions.
Use when auditing Claude skills and commands for quality. Supports Quick Scan (changed skills only) and Full Stocktake modes with sequential subagent batch evaluation.
A comprehensive verification system for Claude Code sessions.
Visualize whether skills, rules, and agent definitions are actually followed — auto-generates scenarios at 3 prompt strictness levels, runs agents, classifies behavioral sequences, and reports compliance rates with full tool call timelines
Validate Git-Ape CLI tool installation (az, gh, jq, git), versions, and auth sessions. Shows platform-specific install commands for anything missing. USE FOR: check Git-Ape prerequisites, what do I need to install for Git-Ape, verify Git-Ape CLI tools, az: command not found, gh: command not found, jq: command not found, git: command not found, az missing, gh missing, jq missing, git missing, fresh machine setup for Git-Ape, dev container setup for Git-Ape, before running git-ape-onboarding, az login required, gh auth login, auth expired, not logged in, outdated az version, minimum az version, upgrade az. DO NOT USE FOR: Anything else. This skill is narrowly scoped to prerequisites checks for Git-Ape's CLI tools and auth sessions. Do not use it for any other purpose.
行为面试 skill。帮求职者挖掘真实经历、用 STAR/CAR 结构化、映射能力标签、构建可复用的中英双语故事库（Story Bank）；并能接入 JD + 简历，针对具体岗位生成 Top 20 BQ 选题 + 基于真实经历的 STAR 准备模板（HTML 报告）。不是背答案，而是建立可复用的职业叙事体系，让任何行为面试题都能自然作答。关键词：behavioral question, BQ, 行为面试, STAR, 故事库, Amazon LP, 职业故事, tell me about a time, 面试准备, JD 面试题预测, top 20 题。
求职工具包。把「看到心动岗位 → 拿到 offer」拆成三个独立子 skill：① Job Description Skill 解码 JD 并出一份 Offer Strategy Report（该不该投 / 匹配度 / 面试题预测），② Resume Skill 把简历结构化并渲染成 11 套打印级模板，③ BQ Skill 挖掘真实经历、STAR 化，建可复用的行为面试故事库。任何求职相关请求（该不该投、改简历、准备行为面试、看 JD、Tell me about a time…）都从这里进，再路由到对应子 skill。关键词：求职, offer, career, job hunt, JD, job description, 简历, resume, CV, 行为面试, behavioral interview, BQ, STAR, tell me about a time, 面试准备。
Build multi-agent AI systems for construction estimation. Use CrewAI/LangGraph to orchestrate specialized agents: QTO agent, pricing agent, validation agent. Automate complex estimation workflows.
Automatically extract patterns, best practices, and reusable knowledge from construction automation sessions to improve future performance.
Detect installed AI coding agents and append honest-feedback rules to their instruction files. Scans for config files (CLAUDE.md, copilot-instructions.md, .cursorrules, etc.), adapts output format per agent, and appends directives that disable sycophancy and enable constructive pushback. Use when setting up honest feedback, disabling people-pleasing, or enabling objective criticism. Triggers on honest agent, objective feedback, no sycophancy, honest criticism, contradict me, challenge assumptions, honest mode, brutal honesty.
What this skill does and when to use it. Claude reads this to decide relevance. Include keywords users would naturally say.
Builds new Claude Code skills with consistent structure, enforced standards, and project-aware configuration. Use when creating a new skill, when the user describes a workflow they want automated, or when the user says they want a new slash command.
Builds new Claude Code agents with consistent structure, enforced standards, and project-aware configuration. Use when creating a new agent, when the user describes a specialised role they want delegated to, or when discussing team composition.
Analyze Chinese notes, articles, manuscripts, and digital card-box material with FIRE 2.0: Full-D numbering, Index keyword webs, Route thinking paths, and Evolution over time. Use when Codex needs to prepare material for semantic search, turn temporary notes into permanent or project notes, build table-of-contents/index/search structures, create compact Chinese FIRE cards, or maintain this skill's original discussion thread provenance.
Create companion Codex workflows from book chapters, reader exercises, prompts, field notes, and teaching material. Use when Codex needs to turn a chapter excerpt, 書籍練習, AI 提示詞, worksheet, or author note into a reusable Skill-style workflow, GitHub-shareable companion artifact, or book link resource for readers.
Use BIRD Book Deconstructor 2.2 to apply the formal BIRD 2.1 Knowledge Address protocol to complex manuscript text and TheBrain Thoughts. Use when splitting books into chapter/section/item knowledge nodes, assigning Book Address, structured Knowledge Index (Weight, Type, Keyword, Alias), Routes, verified Deep Links, and Semantic Roles; producing BIRD Excel workbooks, TheBrain scaffolds, Roam JSON, monochrome printable double nine-grid cards, or a routed handoff from iMandalArt to A4 eight-page booklets; or auditing and migrating existing BIRD/TheBrain indexes.
Aqua is a CLI-first message tool for AI agents. Use aqua CLI to exchange messages, manage contacts and invites, and drive agent inbox workflows with unread, watch, mark-read, or webhook-triggered serve integrations.
Programmatic canvas toolkit for creating, editing, and refining Excalidraw diagrams via MCP tools with real-time canvas sync. Use when an agent needs to (1) draw or lay out diagrams on a live canvas, (2) iteratively refine diagrams using describe_scene and get_canvas_screenshot to see its own work, (3) export/import .excalidraw files or PNG/SVG images, (4) save/restore canvas snapshots, (5) convert Mermaid to Excalidraw, or (6) perform element-level CRUD, alignment, distribution, grouping, duplication, and locking. Requires a running canvas server (EXPRESS_SERVER_URL, default http://localhost:3000).
> VRChat skill renovator for knowledge fill, refresh, and quality improvement. Use this skill when updating VRChat skills to new SDK versions, filling missing knowledge, fixing outdated information, or improving skill quality. Targets unity-vrc-udon-sharp and unity-vrc-world-sdk-3 skills. information audit, catch-up, renovate, refresh, improve skills, SDK update.
A test skill to verify all plugin tools work correctly - use_skill, read_skill_file, run_skill_script, find_skills
| Open Policy Agent integration. Manage data, records, and automate workflows. Use when the user wants to interact with Open Policy Agent data.
| Sessions integration. Manage Sessions, Persons, Organizations, Notes, Files. Use when the user wants to interact with Sessions data.
| SkillAlchemy — 一念落地，万象成形。输入任意想法或蒸馏目标，输出可安装的 SKILL.md。 内部编排 Lens（看清问题）和 LEAP（执行蒸馏/融合）。用户唯一入口。 Use when 用户说「蒸馏」「生成 skill」「融合」「我想做 X 但不知道从哪下手」。
| LEAP — 落地执行引擎。内含两条管线：A 分支蒸馏（从 raw data 提取 skill）、 B 分支融合（多 skill 编织为一个）。被 SkillAlchemy 编排器调用。 Use when 编排器判断需要蒸馏或融合时。
>- Universal discipline for any LM-driven loop — agent retries, plan-act-observe, multi-agent handoffs, optimiser passes, test-fix cycles. Encodes the one rule every framework reliable terminator. Termination must be provided by an explicit counter + exit predicate + stagnation signal + escalation path that live OUTSIDE the LM's control. This is a tool- level, framework-agnostic skill. It maps onto LangGraph (recursion_limit + state counter + interrupt), CrewAI (max_iter + max_rpm + human_input), Claude / OpenAI SDKs (max_iterations + tool_use_budget), DSPy (declared evaluation budget), Aider (REPL + loop, recursion limit, recursion_limit, GraphRecursionError, max iterations, max_iter, agent stuck, agent won't stop, runaway agent, ReAct loop not terminating, agent repeating itself.
>- Cross-framework enhancement overlay for choosing a multi-agent topology BEFORE writing any agent. A binary-question rubric — is single-agent + tools enough? do agents need to know about each other? does the output need one voice? — maps the answer to single-agent / supervisor / swarm / sequential / hierarchical. Activates when a coder agent is tempted to "split the work into roles" or reaches for a multi-agent framework. Encodes the *selection use multi-agent, single vs multi agent, do I need multiple agents, supervisor vs swarm, multi-agent vs single agent, agent team design.
SOP for building multi-agent systems with CrewAI — role-based collaboration, sequential/hierarchical processes, Flows, memory, delegation. Use when modeling agent teams with clear roles and task pipelines.
| Operating SOP for DSPy (Stanford NLP) — the declarative framework for "programming, not prompting" language models. "BootstrapFewShot", "GEPA", "Signatures + Modules", "teleprompter", "auto-tune prompts for a different LM", or whenever a brittle hand-crafted prompt pipeline needs to be turned into a *compiled*, measurable, swappable program. Do NOT activate for one-shot prompt tweaks, no-metric exploratory work, or pipelines where prompts must remain human-authored verbatim — use raw prompting or LangChain templates instead.
| Decision protocol for the map-reduce / dynamic fan-out pattern in LM pipelines — "given list L, run f(item) for each item in parallel, then combine". Activates when the coder agent is about to process N items with N LM calls (per-doc summarize, per-query retrieve, per-candidate rank, parallel tool fan-out). Encodes the *when*, *how many at once*, *what to do when one fails*, and *how to reduce* — not the API of any single Flow, `asyncio.gather`, `ThreadPoolExecutor`, LlamaIndex batch retrieval.
| the framework sent to the model, before changing anything else. Activate when an LM call produced an unexpected output (wrong answer, schema violation, refusal, truncation, cost spike, latency spike, infinite loop, "model got dumber after upgrade"). The skill enforces a 30-second inspect step BEFORE any prompt edit, model swap, retry, or temperature CrewAI `step_callback`, LangChain `set_debug`/`set_verbose`, Aider `/diff`+`--verbose`, raw OpenAI/Anthropic via `OPENAI_LOG=debug`/`ANTHROPIC_LOG=debug` or HTTPX event hooks. Do NOT activate for first-time prompt authoring, exploratory prompt design, or non-LM bugs.
The compile-readiness gate for prompt auto-optimization. Decide whether you have earned the right to run an optimizer (DSPy MIPROv2 / GEPA / BootstrapFewShot) before spending compute. Two preconditions only — a real metric, and enough examples for the optimizer you picked. Garbage metric in, garbage prompt out. Pick the optimizer by data scale; GEPA inverts the scale assumption (~10 examples + textual feedback).
>- Enhancement-overlay SOP for the reranker stage of a RAG pipeline — the "retrieve wide, rerank narrow" discipline. Activate when a calling agent owns a retrieval pipeline whose the context window is under pressure from too many marginal chunks. Encodes the one non- negotiable insight — a cheap bi-encoder retrieves *wide* for recall, then a more expensive cross-encoder (which reads query + document *together*) reranks *narrow* for precision; keep top-N=20-50, rerank to top-k=3-5. Covers when to add a reranker (and when not to), N-vs-k tuning, model choice (Cohere/Voyage API vs bge-reranker local vs SentenceTransformer cross-encoder), latency/cost budgeting, and the cross-framework mapping (LlamaIndex node postprocessors, LangChain ContextualCompressionRetriever, Cohere/Voyage rerank APIs, local cross-encoders). This is an ENHANCE overlay over the per- framework skills — cross-link `[[llamaindex]]` and `[[agentsop-hybrid-retrieval]]` for the
| when to /clear, when to keep context, and how to detect "context bleed" — the failure mode where stale conversation history biases the model against the current task. Surfaces a discipline that Aider (/clear), Claude Code (/clear), CrewAI (memory=False, re-instantiate), and LangGraph (new thread_id, subgraph isolation) all encode separately but none name as a skill.
>- Enhancement overlay for multi-agent / tool-using coder agents. Encodes the per-agent tool- scoping discipline that role-based frameworks (CrewAI, LangChain) document only as a every agent is a correctness and blast-radius risk. Activates when an agent system has tools AND there is more than one agent (or one agent holding many tools). Treat a tool as a capability grant; scope by least-privilege. ENHANCE overlay — read alongside [[crewai]], tools per agent, least-privilege agent, agent tool access, tool permissions, limit agent