routing
Skill0 starsChanged 40 days agoMITView on GitHub
Choose the right agent-router tier (scout, worker, analyst or judge) before spawning a subagent with the Agent tool. Use whenever you are about to delegate work to a subagent, when a spawn was rejected by the agent-router hook, or when deciding whether a task can go to a cheaper agent. Covers how to size the agent, how to write the brief, and how to record what happened to its output.
13 other projects use a version of this.
Adopted by
Most recently taken up by
- AnyOrgName/php-skillstheir own version, first seen Nov 8, 2026
- github/replicatheir own version, first seen Nov 1, 2026
- darknecrocities/SkillBuildertheir own version, first seen Nov 1, 2026
- ThomasLangbroek/frugaltheir own version, first seen Oct 13, 2026
- zprolab/WhaleKittheir own version, first seen Oct 13, 2026
- ypollak2/llm-routertheir own version, first seen Oct 12, 2026
Used with
Projects with it also use parallelization, prompt-chaining and 1 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
hajekimSkillparallelization
This skill should be used when the user wants to "run tasks in parallel", "concurrent LLM calls", "fan-out and fan-in", "parallel agents", "simultaneous execution", "reduce latency with concurrency", "parallel sub-tasks", "parallel processing", "concurrent agent execution", "run agents simultaneously", "speed up with parallelism", "async agent execution", "batch LLM calls", "parallel workflow", "parallelize agent tasks", or needs to execute multiple independent tasks simultaneously. Also responds to Korean: "병렬 작업 실행", "동시 LLM 호출", "팬아웃 팬인 패턴", "병렬 에이전트", "동시 실행", "동시성으로 지연 시간 감소", "병렬로 실행", "동시에 여러 작업", "병렬 처리", "동시에 돌려줘", "빠르게 병렬로", "여러 작업 동시 실행해줘". Also responds to Japanese: "並列処理", "同時実行", "並列エージェント", "並行LLM呼び出し", "ファンアウト・ファンイン", "複数タスクを同時に実行したい", "処理速度を上げたい", "非同期エージェント", "バッチ処理", "並列で動かしたい" Also responds to Chinese: "并行处理", "并发执行", "并行智能体", "同时调用多个LLM", "扇出扇入模式", "多任务并行运行", "加快处理速度", "异步智能体", "批量LLM调用", "同时运行多个任务".. Apply this skill to design or implement the Parallelization agentic design pattern.
Also in 3 projects that use this
- hajekimSkill
prompt-chaining
This skill should be used when the user wants to "chain prompts", "build a pipeline", "break down complex tasks into steps", "sequential LLM calls", "multi-step reasoning", "pipeline pattern", "sequential agent pipeline", "multi-step prompt pipeline", "LLM chain", "step-by-step agent", "prompt pipeline", "decompose task into prompts", "stage-by-stage processing", "chained LLM calls", or when a task is too complex for a single prompt and needs to be decomposed into sequential sub-tasks. Also responds to Korean: "프롬프트 체이닝", "체인 방식 에이전트", "단계별 LLM 호출", "복잡한 작업 분해", "순차적 프롬프트 파이프라인", "단계적 작업 처리", "단계별 처리", "순차 처리 에이전트", "파이프라인 만들어줘", "단계적으로 처리해줘", "연속 프롬프트", "체이닝 해줘". Also responds to Japanese: "プロンプトチェーニング", "ステップごとに処理", "順次LLM呼び出し", "パイプライン構築", "複雑なタスクを分解", "段階的エージェント", "連続プロンプト", "LLMパイプライン", "順番に処理してほしい", "チェーン方式エージェント" Also responds to Chinese: "提示词链", "顺序调用LLM", "构建流水线", "分解复杂任务", "逐步处理", "链式提示", "多步骤流水线", "按顺序处理任务", "提示词管道", "帮我一步一步处理".. Apply this skill to design, implement, or explain the Prompt Chaining agentic design pattern.
Also in 3 projects that use this
NVIDIASkilltesting
Run Python and Rust tests for physicsnemo-curator using uv + pytest with coverage reporting via pytest-cov, and cargo-nextest for Rust tests. Includes benchmark workflows with pytest-benchmark and criterion.
Also in 3 projects that use this
Other versions13 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 13
- github/replica★ 230same name
- ypollak2/llm-router★ 123same name
- ThomasLangbroek/frugal★ 37same name
- darknecrocities/SkillBuilder★ 13same name
- flpbalada/fb-skills★ 8same name
- AnyOrgName/php-skills★ 4same name
- hajekim/agentic-design-patterns-skills★ 4same name
- jkaninda/okapi-skills★ 3same name
- ffsshhttiikk/opencode-agents-skills★ 2same name
- EmmittJ/guild★ 1same name
3 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
8 rules. 0 are common, 6 are distinctive, and 2 are very specific to this project.
0
Common
Most of the 12 edited versions we read keep these.
6
Distinctive
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 8
Every spawn should name a tier.
- Distinctive: line 25
Ask two questions about the task.
- Distinctive: line 27
What does a wrong answer cost?: If a wrong answer would be acted on, merged or shown to the user as true, the cost is high.
- Distinctive: line 34
Start from the tier whose "takes" column matches the work.
- Distinctive: line 35
If the result is cheap to check and you will check it, go down one tier.
- Distinctive: line 38
When unsure between two tiers, take the higher one.
- Very specific: line 40
Do not set model on the call to make a tier cheaper.
- Very specific: line 48
When you send work to judge that another agent produced, include a line Reviews: <agent-id> with that agent's id from its result.
“Common” is measured: kept by at least 6 in 10 of the 12 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Routing a subagent
Every spawn should name a tier. An agent that is too small is the costlier mistake: its work gets redone or, worse, trusted. An agent that is too large only wastes usage.
The tiers
| Tier | Agent type | Takes | Model, effort |
| scout | agent-router:scout | Lookups, file finding, status reads, bounded extraction | Haiku, low |
| worker | agent-router:worker | Implementation with a checkable result (tests, a diff) | Sonnet, medium |
| analyst | agent-router:analyst | Research and synthesis across many sources | Sonnet, high |
| judge | agent-router:judge | Review, adversarial critique, security or design calls | Opus, high |
A project can rename or resize tiers. If the session start message lists a tier table for this project, use the agent types it names; they replace the ones above.
How to choose
Ask two questions about the task.
- What does a wrong answer cost? If a wrong answer would be acted on, merged or shown to the user as true, the cost is high.
- How cheaply can the result be checked? A file path, a test run or a diff is cheap to check. A review verdict, a design recommendation or a security assessment is not.
Then:
- Start from the tier whose "takes" column matches the work.
- If the result is cheap to check and you will check it, go down one tier.
- Judgment work never goes down. Review, audit, critique, security and design calls go to judge even when they look small.
- When unsure between two tiers, take the higher one.
Do not set model on the call to make a tier cheaper. The hook rejects an override below the
tier's floor. An override above it is allowed; use it only when this one task needs it.
Writing the brief
descriptionis logged as the task summary. Make it a short, general label ("find config loader", "review auth diff") with no names, secrets or pasted content.- Say what done means and how the result will be checked.
- When you send work to judge that another agent produced, include a line
Reviews: <agent-id>with that agent's id from its result. This links the verdict to the spawn it judges.
If the hook rejects a spawn
The rejection reason names the call to make instead. Re-spawn as it says. Moving up a tier is always allowed, so a rejection never blocks the work.
Recording outcomes
The log only learns what happened to an agent's output if it is told. When the session start
message gives a mark command, run it in these cases, with the agent's id:
redone: you discarded the agent's result and did the task yourself or re-spawned it.review-failed: a review found a serious problem in that agent's output.user-rejected: the user corrected or reverted that agent's result.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.