Claude-skill-registry intent-first

Clarify user intent before executing tasks. Use when requests are vague, ambiguous, or could be interpreted multiple ways. Helps prevent wasted effort from misunderstanding.

install
source · Clone the upstream repo
git clone https://github.com/majiayu000/claude-skill-registry
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/majiayu000/claude-skill-registry "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/data/intent-first" ~/.claude/skills/majiayu000-claude-skill-registry-intent-first && rm -rf "$T"
manifest: skills/data/intent-first/SKILL.md
source content

Intent-First: Clarify Before You Execute

Purpose

Transform vague requests into clear, actionable specifications before starting work.

When to Use

  • User request is ambiguous or could mean multiple things
  • Task scope is unclear
  • Success criteria are not defined
  • You're about to start significant work

The Intent Clarification Framework

Step 1: Identify the Intent Type

TypeSignal WordsExample
Createmake, build, write, generate"Make a report"
Modifychange, update, fix, improve"Fix this code"
Analyzereview, check, find, understand"Review this design"
Decideshould I, which is better, compare"Should I use X or Y?"
Learnexplain, how does, what is"How does this work?"

Step 2: Ask the 5 Clarifying Questions

Not all questions apply to every request. Select the relevant ones:

1. PURPOSE (Why)

"What problem are you trying to solve?" or "Who will use this and for what?"

2. SCOPE (What)

"What should be included? What should NOT be included?"

3. FORMAT (How)

"What format do you expect? (bullet points, prose, code, diagram)"

4. CONSTRAINTS (Limits)

"Any constraints? (time, length, technology, style)"

5. SUCCESS (Done)

"How will you know this is successful?"

Step 3: Confirm Understanding

Before proceeding, summarize:

I understand you want to:
- [Action]: [specific task]
- [Purpose]: [why this matters]
- [Output]: [expected deliverable]
- [Constraints]: [limitations to respect]

Shall I proceed?

Examples

Vague Request

"Help me with my presentation"

Clarified Intent

I understand you want to:
- Action: Improve/create presentation slides
- Purpose: [Need to ask - what's the presentation for?]
- Output: [Need to ask - slides? script? both?]
- Constraints: [Need to ask - time limit? audience?]

Let me ask:
1. What is the presentation about and who is the audience?
2. Do you have existing slides to improve, or starting from scratch?
3. What do you need help with - content, structure, or visuals?

Anti-Patterns

Don't Do ThisDo This Instead
Start working immediately on vague requestsPause and clarify first
Ask all 5 questions every timeSelect relevant questions only
Make assumptions silentlyState assumptions and confirm
Over-clarify simple requestsUse judgment on when to clarify

Quick Decision: When to Skip Clarification

Skip if ALL are true:

  • Request is specific and unambiguous
  • Scope is clearly bounded
  • Failure cost is low (easy to redo)
  • You have high confidence in understanding

Integration with Other Skills

After clarifying intent:

  • Use delegation-triggers to decide who should execute
  • Use skill-activator to find relevant skills for the task