Loading...
Loading...
Found 13,349 Skills
在新机器或新项目上落地「高智商领导 + 便宜执行」分层子代理：Codex Sol 领导 + Luna 工人， 可选 Claude Code 项目级 agents 与 Pi/pi-flow 跨工具编排。用于： (1) 从零安装并配置 Codex / Claude Code / Pi (2) 写入项目级 .codex/agents、AGENTS.md、.claude/agents (3) 修复 Sol 无法 spawn Luna 的 multi-agent catalog 问题 (4) 跑 Sol/Luna/多代理冒烟验证 触发：新机器设置、Sol-Luna、分层子代理、multi-agent 配置、codex agents 初始化
Interactive session to craft a system prompt for an AI agent powered by the Sanity Context MCP server. Use this skill when users want to define agent personality, set tone/voice, establish boundaries and guardrails, configure refusal behaviors, or control how their agent communicates. Trigger when users mention their agent sounds wrong, needs to refuse certain requests, has the wrong tone, or they want to set communication rules and policies.
Use when a developer wants to iterate on ONE specific Agent Observability / LLM Obs trace whose output they didn't like — re-running that trace against their LOCAL code, seeing a concise diff of the old vs new output, and looping (change code → replay → diff) until satisfied. Invoked as /agent-observability-replay-trace <trace-id> [changes to test]. Signals: "replay this trace"; "iterate on a trace"; "this trace's output is wrong, fix it and re-run"; "re-run trace <id> with <change>"; pasting a trace id from the Agent Observability UI with a description of what to fix. It fetches the trace via the datadog-llmo MCP or the pup CLI, edits code, re-runs the app to emit a NEW trace, and diffs the two — no local server, no browser. For agents traced with ddtrace / LLM Obs (Python first-class), with JSON-serializable entry input. Do NOT use for: scored Experiments or the browser "Replay" button (that's agent-observability-replay-experiment), building an experiment from a dataset/CSV, writing evaluators, root-causing failed traces, or RUM/HTTP session replay.
OpenTelemetry in Java — Javaagent zero-code instrumentation, Spring Boot Starter, manual autoconfigure SDK, declarative YAML configuration, BOM dependency management, sensitive-data capture and redaction (url.query, headers, request parameters, SQL sanitization). Use when adding, reviewing, or configuring OpenTelemetry in a Java service. Triggers on "setup otel in java", "java telemetry", "javaagent", "Spring Boot otel", "GlobalOpenTelemetry", "AutoConfiguredOpenTelemetrySdk", "TracerProvider java", "url.query redaction", "capture request headers", or any Java-related OTel question.
Design lightweight eval harnesses for coding or workflow agents, especially when testing whether agents follow steering docs, use tools correctly, preserve user edits, recover from failures, and produce the intended artifact.
Use when an agent is asked "what did I (or my team) work on yesterday / this week / today" across provider data in a relayfile mount (Linear, GitHub, Notion, Slack, Confluence, Jira, etc.). Tells the agent to consult the pre-computed `digests/yesterday.md` (and sibling digest files) at the workspace root BEFORE doing manual exploration with `ls`/`grep`/`find`. The digest is deterministic, exhaustive over the window, and costs one file read instead of dozens of provider queries.
Deterministic issue-relationship graph over GitHub native sub-issues + dependencies — compute the ready set / parent-rollup candidates / close-kick targets as pure calculation (scripts, no LLM judgment), write real edges when creating spin-off issues, and mutually exclude terminal actions across parallel agents via claim-comment fencing. Called by issue-sweep (candidate injection), issue-review (edge writing + rollup), and the future agent:ready producer routine.
Meta-router and multi-agent conductor for design work that needs skill selection or a multi-stage pipeline — visual frontend (web pages, landing pages, product UI, mobile screens), data visualization, HTML deliverables (reports/diagrams/plans), artifacts, motion polish, and module/API interface design. The STABLE unified entry point — member skills churn underneath, this router discovers them live, picks one direction authority, delegates image-generation stages to Codex workers, and closes every pipeline with an evidence-first anti-slop quality loop. Do NOT invoke for a one-line CSS/copy tweak, for backend-only work, or when the user explicitly names a single member skill for a single-skill-sized task — those go direct.
Fill-in-the-blank prompt templates for delegating work to subagents or tmux workers - SEARCH/LOCATE, IMPLEMENT, REFACTOR, RESEARCH, REVIEW/VERIFY. Invoke BEFORE writing any delegation prompt (Agent tool, agent-tmux worker, or fanout task) so the prompt ships with explicit GOAL, ACCEPTANCE, and REPORT sections instead of a vague ask. Not for deciding WHETHER to delegate or for driving workers after launch (see the tmux-delegate agent and the tmux-agent-tools skill for those).
Use when parallel agents share a codebase: adding new behavior, editing or resolving conflicts in shared files (dispatchers, registries, lockfiles), finding and splitting churn hotspots, or writing AGENTS.md, CLAUDE.md, or README.
Enable a symmetric collaborative closed-loop of "Primary Executor + Independent Reviewer" between Codex and Claude Code via native CLI. This applies when users require the two models to work together, cross-validate, conduct independent reviews, revise until ACK, or when tasks involve product requirement convergence, complex solutions, cross-module development, migration, security, major refactoring, and high-quality delivery; regardless of whether initiated from Codex or Claude Code, the current model will take charge and fully invoke the other party to complete requirement challenge, scheme gatekeeping, implementation cold review, and final acceptance.
Use when an agent needs to drive the full agent-manager lifecycle through `amctl` — install the CLI, log in, create/deploy an agent, list projects and agents, watch build progress, fetch build/runtime logs and metrics, and pull traces.