mcpField
V
vague

vague

github.com/team-attentionAgent workflows82837 installsUpdated 27w ago
Repository

About vague

vague is a skill for turning ambiguous requests into actionable specifications through hypothesis-based questions. It requires AskUserQuestion and presents a before-and-after summary once key uncertainties are resolved.

Key features

  1. 01Identify ambiguities in requirements
  2. 02Frame questions as hypotheses
  3. 03Batch up to four related questions
  4. 04Summarize requirements before and after
  5. 05Ask whether to save the result

Use cases

  1. 01Clarify a vague feature request
  2. 02Fill gaps in an incomplete bug report
  3. 03Specify an underspecified task

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/team-attention/plugins-for-claude-natives --skill vague

After installing, type/vaguein Claude Code to confirm it loaded.

Details

From the project README

Vague: Requirement Clarification

Transform vague or ambiguous requirements into precise, actionable specifications through hypothesis-driven questioning. ALWAYS use the AskUserQuestion tool — never ask clarifying questions in plain text.

When to Use

  • Ambiguous feature requests ("add a login feature")
  • Incomplete bug reports ("the export is broken")
  • Underspecified tasks ("make the app faster")

For strategy/planning blind spot analysis, use the unknown skill. For content-vs-form reframing, use the metamedium skill.

Core Principle: Hypotheses as Options

Present plausible interpretations as options instead of asking open questions. Each option is a testable hypothesis about what the user actually means.

BAD:  "What kind of login do you want?"           ← open question, high cognitive load
GOOD: "OAuth / Email+Password / SSO / Magic link" ← pick one, lower load

Protocol

Phase 1: Capture and Diagnose

Record the original requirement verbatim. Identify ambiguities:

  • What is unclear or underspecified?
  • What assumptions would need to be made?
  • What decisions are left to interpretation?
Phase 2: Iterative Clarification

Use AskUserQuestion to resolve ambiguities. Batch up to 4 related questions per call. Each option is a hypothesis about what the user means.

Cap: 5-8 total questions. Stop when all critical ambiguities are resolved, OR user indicates "good enough", OR cap reached.

Example AskUserQuestion call:

questions:
  - question: "Which authentication method should the login use?"
    header: "Auth method"
    options:
      - label: "Email + Password"
        description: "Traditional signup with email verification"
      - label: "OAuth (Google/GitHub)"
        description: "Delegated auth, no password management needed"
      - label: "Magic link"
        description: "Passwordless email-based login"
    multiSelect: false
  - question: "What should happen after registration?"
    header: "Post-signup"
    options:
      - label: "Immediate access"
        description: "User can use the app right away"
      - label: "Email verification first"
        description: "Must confirm email before access"
    multiSelect: false
Phase 3: Before/After Summary

Present the transformation:

## Requirement Clarification Summary

### Before (Original)
"{original request verbatim}"

### After (Clarified)
**Goal**: [precise description]
**Scope**: [included and excluded]
**Constraints**: [limitations, preferences]
**Success Criteria**: [how to know when done]

**Decisions Made**:
| Question | Decision |
|----------|----------|
| [ambiguity 1] | [chosen option] |
Phase 4: Save Option

Ask whether to save the clarified requirement to a file. Default location: requirements/ or project-appropriate directory.

Ambiguity Categories

CategoryExample Hypotheses
ScopeAll users / Admins only / Specific roles
BehaviorFail silently / Show error / Auto-retry
InterfaceREST API / GraphQL / CLI
DataJSON / CSV / Both
Constraints<100ms / <1s / No requirement
PriorityMust-have / Nice-to-have / Future

Rules

  1. Hypotheses, not open questions: Every option is a plausible interpretation
  2. No assumptions: Ask, don't assume
  3. Preserve intent: Refine, don't redirect
  4. 5-8 questions max: Beyond this is fatigue
  5. Batch related questions: Up to 4 per AskUserQuestion call
  6. Track changes: Always show before/after

Comments

Markdown supported