clarifying-requirements
Skill11 starsChanged 50 days agoMITView on GitHub
Analyzes user prompts for gaps, ambiguities, and unstated assumptions, then asks clarifying questions before work begins. Use when requests are vague, lack acceptance criteria, or have hidden complexity.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- yuchi-chang/no-capetheir own version, first seen Oct 8, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- yuchi-chang/no-cape★ 2same name
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.
In this repository44 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- auditing-uiSkill
- browser-testingSkill
- capturing-product-mediaSkill
- ci-local-exportSkill
- ci-pulseSkill
- ci-quarantineSkill
- ci-statusSkill
- creating-pull-requestsSkill
- debugging-systematicallySkill
- debugging-test-failuresSkill
- debugging-typescript-errorsSkill
- deps-weeklySkill
- designing-frontendSkill
- opensrcSkill
- organizing-fsd-architectureSkill
- runtime-watchSkill
- stewarding-ci-pipelineSkill
- syncing-agent-skillsSkill
- verification-before-completionSkill
- vision-drift-checkSkill
- visual-companionSkill
- writing-adrsSkill
- writing-app-copySkill
- writing-blog-postsSkill
- writing-changelogsSkill
- writing-developer-guidesSkill
- writing-for-humansSkill
- writing-to-usersSkill
- adding-config-fieldsSkill
- maintaining-dependenciesSkill
- maintaining-dev-playgroundSkill
- managing-specsSkill
- marketplace-devSkill
- orchestrating-parallel-workSkill
- reading-session-transcriptsSkill
- receiving-code-reviewSkill
- requesting-code-reviewSkill
- styling-with-tailwind-shadcnSkill
- test-driven-developmentSkill
- upgrading-runtime-dependenciesSkill
- working-in-worktreesSkill
- dorkbot-triageSkill
The file
We split this skill into the rules it gives an agent
5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈4
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 38
Ask for information you can find — search the codebase first
- Distinctive: line 39
Ask questions that don't change what you'd build
- Distinctive: line 110
Don't explore blindly. Ask questions first, then explore with purpose.
- Very specific: line 114
If a request is vague, clarify BEFORE running ideation (/flow:ideate, from the installed flow plugin) — ideation works better with clear …
- Distinctive: line 118
Not everything needs formal workflows.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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
Clarifying Requirements
Analyze user prompts for gaps, ambiguities, and unstated assumptions — then ask the questions the user failed to ask BEFORE beginning work.
Core Principle
Don't just answer what was asked — anticipate what SHOULD have been asked. Users often don't know what they don't know. Surface gaps in their thinking, unstated assumptions that could derail implementation, and scope ambiguities that lead to rework. The one test that matters: "Can I implement this without making assumptions?" If not, ask about the assumptions you'd have to make.
When to Apply
Activate proactive clarification when the request shows:
| Signal | Example | Why It Matters |
| Vague action verbs | "add", "improve", "fix" without specifics | Undefined scope leads to wrong implementation |
| Missing constraints | "make it faster" without metrics | No way to know when you're done |
| Complexity underestimation | "just add a button", "simple form", "quick" | Often hides error/loading/permission states |
| Multiple features bundled | "add X and Y and also Z" | Scope creep, unclear priority |
| Goal without criteria | "users should be able to..." | No acceptance criteria |
| Assumed context | "fix the bug" (which bug?), "like the other one" | Missing reproduction steps / referent |
Questioning Strategy
Do:
- Limit to 2-4 questions — pick the highest-impact clarifications
- Explain why each question matters ("This affects routing, state management, and URL structure")
- Group related questions by theme (scope vs technical decisions)
- Use AskUserQuestion for bounded choices — see Recommendation Discipline below
Don't:
- Ask for information you can find — search the codebase first
- Ask questions that don't change what you'd build
- Repeat questions already answered in the conversation
- Delay genuinely simple, clear tasks
- Ask without recommending — never present options without a position (rare exception: pure aesthetic preferences with no codebase precedent)
Recommendation Discipline
Core principle: Asking a question without a recommendation is lazy.
You have the codebase, research findings, existing patterns, and domain knowledge. Use them to form a point of view on every question. The user hired you to think, not just to enumerate options.
Why it matters: it saves the user cognitive load (they evaluate your reasoning instead of starting from scratch), proves you've thought about the problem in context, turns a back-and-forth into a single confirmation, and forces you to notice when an option is clearly wrong.
How to Form a Recommendation
- Codebase precedent — "The codebase uses X in 4 places" is strong evidence
- Research findings — a revealed best practice carries weight
- Architectural fit — which option aligns with the project's conventions?
- Simplicity — when in doubt, recommend the simpler option; complexity needs justification
- Blast radius — prefer options that touch fewer files and introduce fewer new patterns
AskUserQuestion Format
The first option MUST be your recommendation, marked (Recommended) in the label. Descriptions explain WHY, based on evidence:
AskUserQuestion:
questions:
- question: "Where should the new settings panel live?"
header: "Location"
options:
- label: "Extend existing SettingsDialog (Recommended)"
description: "Follows FSD patterns — SettingsDialog already handles config. Adding a tab keeps navigation consistent with the existing 3-tab layout."
- label: "New standalone panel"
description: "More isolation but introduces a new navigation pattern. Would need its own route and state management."
- label: "Inline in the sidebar"
description: "Quick access but limited space. Would need collapsible sections for complex settings."Anti-Pattern: The Option Menu
❌ BAD — No recommendation, no reasoning:
"How should we store this data?"
- Option A: localStorage
- Option B: Database
- Option C: File system
✅ GOOD — Clear recommendation with evidence:
"How should we store this data?"
- localStorage (Recommended) — Matches the existing theme/preference storage pattern.
Single-user tool, no cross-device sync needed.
- Database — Overkill for user preferences. Would add a migration and schema change
for data that's inherently client-scoped.
- File system — Only useful if the CLI needs access. Currently no CLI preference commands exist.Integration with Existing Workflows
With Plan Mode
Plan mode has a dedicated clarification phase — apply this skill BEFORE exploring:
❌ Wrong order:
EnterPlanMode → Explore everything → Ask questions → Re-explore
✅ Right order:
EnterPlanMode → Ask key questions → Explore with focus → Present planDon't explore blindly. Ask questions first, then explore with purpose.
Before Ideation and Specification
If a request is vague, clarify BEFORE running ideation (/flow:ideate, from the installed flow plugin) — ideation works better with clear inputs. Specifications especially must not be built on assumptions.
In Regular Conversation
Not everything needs formal workflows. For casual requests that trigger the signals above, ask clarifying questions inline before responding.
Balancing Act
The goal is to be helpful, not interrogative:
| Situation | Approach |
| Simple, clear request | Just do it |
| Mostly clear, one ambiguity | Ask one focused question |
| Multiple ambiguities | Ask 2-3 key questions, note you'll ask more as you go |
| Fundamentally unclear | Pause and clarify before any work |
| User seems rushed | Offer to proceed with stated defaults |
| User is exploring | Match their exploratory energy, don't over-formalize |
Summary
- Analyze every non-trivial request for gaps, ambiguities, assumptions
- Ask the questions users didn't know to ask
- Always recommend — every question gets a recommendation backed by evidence
- Limit to 2-4 high-impact questions per interaction
- Don't delay simple tasks with unnecessary questions
- Surface risks and alternatives proactively
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.

