Agent Workflow Claude Skills
4 122 agent workflow skills from 665 authors. They configure the agents themselves: memory, prompts, context and other skills. Half of them fit into 1 844 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 589 copies of these same skills sitting in other people's repositories — counted once here, not 589 times.
4 122 unique 665 authors 4 updated this month 506 from vendors
An example skill for testing diff-skill evaluation.
An example skill for testing diff-skill evaluation.
> Use when the user has a prompt that feeds a system they can already score, and wants that prompt automatically improved to raise the score against their own evaluation command. Makes one targeted quality edit per iteration — clarity, context, specificity, structure, examples, decomposition, guardrails — re-runs the user's eval to measure the metric, and keeps the edit only if the metric improves, else reverts; loops to a target, plateau, or budget. The metric is whatever the user's eval command prints (task accuracy, an LLM-judge score, a pass rate, a tool-call success rate); the loop is metric-agnostic and never edits the eval. Not for writing a prompt from scratch, not for tuning model weights or hyperparameters, and not for one-off manual prompt edits without a score.
> Use when the user wants to adversarially stress-test a guardrail, classifier, prompt, or API they own or are authorized to test, to surface the distinct ways it fails. Generates adversarial inputs, runs them through the target and a ground-truth oracle, logs every target-vs-oracle disagreement as a failure de-duplicated by technique class, and loops until rounds stop surfacing new classes. Produces a catalogue of distinct, reproducible failures — the attacker half of a find→fix setup. Not for patching the target, and not for attacking systems the user does not own or have permission to test.
> Use when the user has a coding prompt — a feature, bug fix, or refactor — and wants it implemented end to end by a self-checking software loop, not a single pass. It refines the prompt into an subagent implements each PR-sized task editing only source, a separate QA subagent authors the tests that prove the task's acceptance criteria (or grades quality only when tests already cover them), runs the task's tests plus the full accumulated regression suite, and grades the code against a strict conciseness/readability/style-match rubric; the Engineer and QA loop on each task until the tests pass and the quality gate holds, then it commits and opens a pull request for the task — stacked PRs (one per task, each based on the prior) by default, or one combined PR — and moves to the next task. Author and critic have disjoint write scopes (Engineer owns source, QA owns tests) so neither can game the gate. Not for producing a plan without building it (that is plan-loop), and not for tuning a metric on an existing artifact under a fixed correctness bound (that is optimize-loop).
> Use when the user wants an autonomous ML research loop that pressure-tests competing ideas before spending compute — several research subagents each propose one architecture change, a self-calibrating Judge critiques them against a rubric, the proposers refine, and the Judge picks the single change to run. The Judge learns to pick better over time by scoring its own predictions against realized metric deltas, recording predicted-vs-realized in a calibration ledger and refining its working rubric. The result is an experiment ledger where each iteration's change won a de-biased tournament. Not for running a single pre-decided experiment, and not for analysis-only exploration — for one hypothesis proposed and run per iteration without competition, use the sibling ml-autoresearch loop.
Analyze LLM experiment results. Handles single or comparative experiments, exploratory or Q&A modes. Use when user says "analyze experiment", "compare experiments", "analyze against baseline", or provides one or two experiment IDs for analysis.
Datadog skills for AI agents. Essential monitoring, logging, tracing and observability.
End-to-end Agent Observability pipeline for an instrumented ml_app — classify production traces, root-cause failures, bootstrap evaluators, then (optionally) sample + publish a dataset, generate + run an experiment, and analyze results. Six narrated phases with a standardized banner and a "continue" checkpoint between each. Pure orchestration over the agent-observability sub-skills (`agent-observability-session-classify`, `agent-observability-trace-rca`, `agent-observability-eval-bootstrap`, `agent-observability-experiment-py-bootstrap`, `agent-observability-experiment-analyzer`). Use when user says "run the eval pipeline", "go from traces to evals", "bootstrap evals end to end", "classify then RCA then bootstrap", "build an eval set from scratch", "onboard me to datasets and experiments", "walk me through experiments", "I have an ml_app, now what", "Agent Observability onboarding", "guided experiment setup", "from traces to experiments", or wants a deterministic, narrated tour from production data through evaluators, datasets, and experiments. Stop early with `--stop-after <phase>` to short-circuit at evaluators or dataset, or resume mid-flow with `--start-at <phase>`.
> Classify whether user intent was satisfied in a Datadog Agent Observability trace or session. (2) trace_id — classify a single Agent Observability trace without RUM; (3) ml_app — sample and classify multiple sessions or traces from a given LLM app. Output is compact by default (verdict + one-sentence reason). Use when evaluating satisfaction, classifying sessions/traces, labeling data, or generating signal for agent-observability-eval-pipeline or agent-observability-trace-rca.
>- Run an iterative code-improvement hill-climb against real Datadog LLM-Obs data, locally, with Claude Code as the agent. Establishes a baseline eval, makes one focused change, re-scores with the same harness, keeps the change if it improves the score in the goal's direction (labeling within-noise gains tentative), and repeats. Use when the user says "run an auto experiment", "hill-climb this code", "iteratively improve X and measure the delta", "optimize this prompt/file against my traces", "auto-optimize against LLM-Obs", or wants the local equivalent of the auto_experiments worker. Works from a local dataset file, an ml_app, a dataset_id, or a list of trace_ids.
Generate a live Single Step Instrumentation (SSI) onboarding confirmation report — verifies APM instrumentation is working end-to-end with deep links into the Datadog UI. Only use after agent-install and enable-ssi have both completed successfully.
Root cause analysis on production LLM traces. Diagnoses why an LLM application is failing — works from eval judge verdicts, runtime errors, or structural anomalies depending on what signals are present. Walks the span tree from symptom to root cause. Use when user says "what's wrong with my app", "why is my eval failing", "analyze errors", "root cause analysis", "diagnose failures", or wants to understand production failure patterns.
APM - install, onboard, instrument, enable, set up, configure, traces, services, dependencies, performance analysis. Use for any request involving Datadog APM setup, instrumentation (SSI, ddtrace, agent install), or analysis.
Install the Datadog Agent on Linux hosts via SSH with Single Step Instrumentation (SSI) enabled — SSI automatically instruments applications for APM without code changes. Only use if no agent is installed yet.
Generate a live Single Step Instrumentation (SSI) onboarding confirmation report for Linux hosts — verifies APM instrumentation is working end-to-end with deep links into the Datadog UI. Only use after agent-install and enable-ssi have both completed.
Configure Unified Service Tags and verify Single Step Instrumentation (SSI) injection on Linux hosts — SSI automatically instruments applications for APM without code changes. Only use if the Datadog Agent is already installed.
Audit what the Bits AI assistant (MCP server) has done in your Datadog org — tool calls by user, resources accessed, and anomaly flags for AI governance.
Security forensics for git repos, AI skills, and MCP servers. Audits dependencies, detects prompt injection, credential theft, runtime dynamism, manifest drift, known CVEs, CISA KEV (actively exploited) vulns, and 2026 attack patterns. Not for fixing vulnerabilities or pentesting.
| Cross-agent self-inspection of your AI-agent stack. Audits skills, MCP servers, hooks, plugins, commands, credentials, and memory files across Claude Code, Codex, OpenClaw, and NanoClaw. Produces a structured inventory and narrative briefing with cross-ecosystem risk analysis. Use when the user asks to audit their own setup, check what they have installed, review their agent stack security posture, or understand cross-tool interactions. Use when a user has accumulated skills/plugins/MCP servers over time and wants visibility into their attack surface. Use after installing new skills or plugins. Do NOT use for vetting external code before install (that is repo-forensics). Do NOT use for incident response during active attacks. Do NOT use for fixing or patching vulnerabilities (forensify is read-only).
| 当系统提示词需要定义 AI 如何分类用户意图、路由到不同处理流程、决定澄清策略和自主度级别时调用此 Skill。适用于多任务型 AI 助手、客服机器人、编程工具、研究助手等需要结构化对话管理的场景。不适用于：纯问答型系统（无任务执行）、单轮交互（无对话状态）、简单的 prompt 模板（无路由逻辑）。当需求仅涉及"输出什么格式"而非"如何决定输出什么"时，应该用 output-formatting 而非本 Skill。
| 当系统提示词需要设计 token 预算分配、上下文压缩策略、延迟加载机制、记忆持久化方案时调用此 Skill。适用于长对话 AI 助手、代码编辑器集成、研究型 Agent、多会话系统等需要精细管理上下文窗口的场景。不适用于：单轮交互系统（无上下文管理需求）、纯无状态 API（无对话历史）、简单的 prompt 模板设计。当需求聚焦于"如何搜索外部信息"而非"如何管理已有信息"时，应该用 search-integration 而非本 Skill。
| 当系统提示词需要设计多代理协作架构、子代理专业化分工、代理间上下文隔离与传递机制、任务生命周期管理时调用此 Skill。适用于 AI Agent 平台、多工具编排系统、代码审查流水线、跨应用协作场景等。不适用于：单代理系统（无委派需求）、简单工具调用（无子代理概念）、纯 API 编排（无 AI 决策）。当需求聚焦于"单代理内的对话路由"而非"多代理间的任务分配"时，应该用 conversation-flow 而非本 Skill。
| 当需要为 AI 产品定义核心身份、角色声明和能力边界时调用此 skill。典型场景包括：设计新 AI 产品的 system prompt 首段、为不同场景创建差异化角色（如教学助手 vs 编程代理）、重新定义 AI 与用户的关系框架。 不适用于：纯人格风格调优（应使用 personality-system）、安全规则制定（应使用 safety-guardrails）、工具集成（应使用 tool-specification）。 关键 trigger 信号：产品需要 AI 有明确自我认知、多个角色共享底层能力但身份不同、需要限制 AI 在特定领域内运作、用户会问"你是谁"。
> Per-project glossary of key definitions, abbreviations, and command-phrases, stored in `GLOSSARY.md` at the project root. Use this skill when the user defines or asks about a project-specific term — variable names, dataset or database names, acronyms — or sets up a command-phrase (a phrase that maps to an action, e.g. "push" = commit and push the paper to GitHub). Triggers "what does X mean here", "what does X stand for", "from now on X means Y", "show the glossary", "what's in our glossary", and "remove X from glossary". passing, or when you hit an undefined abbreviation or variable name in their code or data. Loaded at session start by `/spin-up` so command-phrases stay active.
> Reference for writing and improving Claude Code skills. Use when the user asks to create a new skill, improve an existing skill, turn a conversation workflow into a reusable skill, or says anything like "make this a skill", "save this as a skill", or "let's capture this."
> Start-of-session orientation routine that briefs Claude on the current state of a project before work begins. Use this skill when the user says "spin up", "spin it up", "let's go", "start up", "spin up the project", "punch it chewy", or any variant signalling they want a session kickoff divergence), CLAUDE.md and README read, most-recent session log read (focusing on "Where we left off"), PINBOARD.md open items, GLOSSARY.md load, and a short synthesis of project state. Ends by asking what to work on today.
> End-of-session cleanup routine that captures all session work before context is lost. Use this skill when the user says "wrap up", "wrap-up", "let's wrap it up", "let's wrap up", "wind down", "end session", "that's it for today", "save and close", "let's call it", "close out", "done for today", or any indication they are finishing a work session and want everything documented before starting fresh.
Run a structured 5-pass finishing audit on any website before launch — scoring visual polish, technical foundation, UX completeness, content quality, and cross-device readiness on 100 points. Use when: **Pre-launch** - Final validation before going live; **Post-redesign** - Verify nothing broke during the overhaul; **Client handoff** - Structured proof that the site is ready; **Quarterly review** - Catch accumulated debt; **Single-pass focus** - Run just Pass 2 after a perf sprint
Use when context is growing large (50k+ tokens), performance is degrading, instructions are being ignored mid-conversation, or planning multi-agent workflows. Triggers on "lost context", forgotten instructions, or sessions exceeding 30 minutes.
Route multi-step marketing, product, or business challenges to the right sequence of ClawFu skills. Recommends which frameworks to combine, in what order, with handoff outputs between steps. Use when: planning a product launch end-to-end, combining positioning + offers + launch skills, running a customer-validation sprint, creating a content strategy pipeline, deciding which skills to chain for a complex project, or orchestrating a sales-enablement playbook.
Process large codebases (>100 files) using the Recursive Language Model pattern. Orchestrates parallel sub-agents to map-reduce across files without context rot. Use when: analyzing large repositories; auditing security or auth across many files; finding patterns across 50+ files; processing large log files or data dumps
Use when testing skills, commands, or agents for quality. Use after creating new skills, before deploying agents, or when debugging inconsistent agent behavior. Triggers on "evaluate", "test quality", "is this skill working", or QA of AI workflows.
Use this skill when you need to parse Word/HTML/JSON/Markdown/Excel requirements and produce a structured analysis; triggers include requirements analysis plus and requirement document parsing.
Use this skill when you need structured test-case review findings from requirements, strategy, and case docs; triggers include test case reviewer plus and advanced test case review.
Use this skill when you need to analyze requirements, identify test points, boundaries, dependencies, and risks before test design; triggers include requirements analysis and test point analysis.
Use this skill when you need a structured test strategy from requirement, analysis, tech, and plan docs; triggers include test strategy plus and advanced test strategy.
Use this skill when you need high-quality test cases from requirements and analysis artifacts; triggers include testcase writer plus and advanced test case writing.
Use this skill when you need to analyze requirements, identify test points, boundaries, dependencies, and risks before test design; triggers include 需求分析 and requirements analysis.
Use this skill when you need to parse Word/HTML/JSON/Markdown/Excel requirements and produce a structured analysis; triggers include 需求分析增强、requirements analysis plus and requirement parsing.
Generates hierarchical context files (CLAUDE.md) throughout a project directory tree, providing AI agents with directory-specific knowledge for better code understanding. Use when setting up a new project for AI-assisted development.
Creates structured context summaries for continuing work across AI sessions. Use when ending a session to enable seamless continuation in a new session without losing context.
> Create, edit, evaluate, and package agent skills. Use when building a new skill from scratch, improving an existing skill, fixing a skill that never triggers or fires unreliably, running evals to test a skill, benchmarking skill performance, optimizing a skill's description, reviewing third-party skills for quality, or packaging skills for distribution — even if the user doesn't explicitly say "skill" (e.g. "teach Claude to do X", "make the agent always follow Y"). Not for using skills or general coding tasks.
The checkpoint coordinator for /feature. Routed to by /feature once a writing-plans plan exists. Groups tasks into batches, delegates each batch to subagent-driven-development (fresh author agent per task, full review chain, fresh verification), then stops for a human checkpoint before the next batch. The checkpointed counterpart to /sprint's autonomous run.
The parallel fan-out primitive. Routed to by any skill or command that splits work across independent units and dispatches an agent per unit — subagent-driven-development, /sprint, parallel /review. It owns the dispatch/collect/funnel discipline: bound concurrency, isolate units, collect every result, dedupe overlap, and funnel through finding-triage then checkpoint-aggregator. Raw agent output is never consumed before the funnel runs; an agent that errors drops its unit without corrupting the batch.
The implementation engine. Routed to by /sprint (full plan, autonomous) and by executing-plans (scoped batch, checkpoint-gated). One fresh subagent per task — test-first via tdd — followed by spec-compliance review, quality review, and fresh-run verification. No single context accumulates drift, and nothing is accepted on a subagent's word.
OPTIONAL per-task isolation for autonomous parallel work. Routed to only on explicit opt-in by subagent-driven-development or dispatching-parallel-agents, so parallel units mutate files without colliding. Stands up one worktree per unit, works it in isolation, then integrates each unit back onto the caller's working branch for the caller's single commit-gate + finishing-a-development-branch exit. Never on the default path; it does not bypass a gate or finish per unit.
The authoring gate for new skills. Routed to when the user invokes /new-skill "<gap>". Five gated phases — gap evidence, scope, authoring, self-review against the v2 house style, routing integration. A new skill is not written until an existing one is proven not to cover the gap, and not shipped until it carries gated phases, hard rules, and a routing entry. Every authored skill matches the v2 format (frontmatter name+description, # name, Pre-flight, Phase N · gate, Hard rules).