request-triage
Skill1 starsChanged 4 months agoMITView on GitHub
触发:新请求需判断处理深度;不触发:已进入澄清→用pm-requirement-intake;价值不清→用pm-value-anchor;输出:流程层级+下一步
In this repository17 other files it gives its agents.
- audience-judgment-framingSkill
- embedded-prd-auditSkill
- freeze-readiness-checkSkill
- memory-export-summarySkill
- pm-demo-designSkill
- pm-design-absorptionSkill
- pm-devil-advocateSkill
- pm-embedded-prdSkill
- pm-prd-writerSkill
- pm-project-opsSkill
- pm-react-prototype-executionSkill
- pm-requirement-intakeSkill
- pm-requirement-reverse-auditSkill
- pm-value-anchorSkill
- prototype-export-intakeSkill
- rule-change-signalSkill
- state-auditSkill
The file
We split this skill into the rules it gives an agent
8 rules. About 0 are common, 6 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
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
Use this skill to choose the right workflow depth before product work begins.
- Distinctive: line 34
Do not use this skill to:
- Very specific: line 58
Use fast-lane when most are true:
- Distinctive: line 60
no new business rule
- Distinctive: line 61
no meaningful state model change
- Distinctive: line 77
do not create a full project workflow unless the change repeats
- Very specific: line 96
Use deep-lane when any high-risk condition is present:
- Distinctive: line 139
Keep it short. The output should help the user decide how much process the request deserves.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Request Triage
Use this skill to choose the right workflow depth before product work begins.
The goal is to prevent two failures:
- over-processing small requests with a heavy workflow
- under-processing high-risk requests that need deeper intake, audit, prototype, or delivery handling
Core Job
Classify the request into one lane:
fast-lane: small, low-risk, clear changestandard-lane: normal product requirement that needs structured intake and a clear handoffdeep-lane: high-risk, cross-surface, rule-heavy, state-heavy, or strategically important requirement
Then recommend the minimum useful next skills.
Use When
Use when:
- a new request arrives and workflow depth is unclear
- the user asks whether a requirement needs full workflow
- a request looks small but may hide rule/state risk
- there are many concurrent requests and prioritization of process depth matters
Do Not Use When
Do not use this skill to:
- decide final business correctness
- write the PRD
- design the demo
- replace
pm-requirement-intakefor unclear requirements - force every request into a formal process
Triage Inputs
Look for:
- change type: copy, config, UI, flow, rule, data, permission, payment, cross-layer behavior
- affected surfaces: backend, App, embedded PRD, data import/export, user-facing display
- user/business impact
- risk if wrong
- number of roles involved
- state and exception complexity
- dependency on existing assets, components, or project baselines
- whether the request is reversible
- whether it needs review, delivery freeze, or only a quick adjustment
Lane Rules
Fast Lane
Use fast-lane when most are true:
- scope is small and clear
- no new business rule
- no meaningful state model change
- low user/business risk
- low dependency on other systems
- rollback is easy
- a short note or direct edit is enough
Typical examples:
- wording change
- copy adjustment
- simple display ordering
- minor field label clarification
- small prototype visual correction
Recommended next move:
- answer directly or make the small change
- optionally record a short changelog line
- do not create a full project workflow unless the change repeats
Standard Lane
Use standard-lane when:
- the request changes page behavior, flow, field rules, or user-facing interaction
- scope is bounded but needs structured clarification
- state and exception coverage matters but is not extreme
- PRD or demo output may be needed
Recommended next skills:
pm-requirement-intakepm-demo-designwhen page/demo expression is neededpm-prd-writerwhen implementation-readable PRD is neededstate-auditbefore freeze or handoffpm-project-opsif artifacts should be saved
Deep Lane
Use deep-lane when any high-risk condition is present:
- cross-surface behavior, such as backend config plus App display
- payment, entitlement, access control, membership, course/content availability, or data migration
- many states, roles, exceptions, or historical compatibility concerns
- unclear success definition or business value
- high coordination cost across product, design, development, testing, operation, or content teams
- likely need for runnable prototype, embedded PRD, or review baseline
- failure would mislead users, create financial risk, or cause significant rework
Recommended next skills:
pm-value-anchorpm-requirement-intakerule-change-signalwhen scope/rule drift appearsstate-auditfreeze-readiness-checkpm-demo-designpm-prd-writerpm-react-prototype-executionwhen runnable behavior matterspm-embedded-prdwhen review-grade linked delivery is neededpm-project-ops
Working Flow
- Restate the request in one sentence.
- Identify risk signals and simplification signals.
- Choose one lane.
- Explain the reason briefly.
- Recommend the minimum next skill path.
- Name what should not be done yet.
Output Format
## Request triage
- Request:
- Lane:
- Why:
- Risk signals:
- Simplification signals:
- Recommended next path:
- Do not do yet:Keep it short. The output should help the user decide how much process the request deserves.
Self-Check
Before finishing, check:
- Did I avoid making the workflow heavier than the request deserves?
- Did I catch hidden risk signals?
- Did I recommend only the next useful skills, not the whole system?
- Did I clearly say what not to do yet?
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.

