← Browse

@charlieviettq/interview-me

A

Extract user intent through structured one-question-at-a-time interviews before specs or implementation. Use when requirements are fuzzy, the user says \"help me think through\", or you need success criteria before planning.

skillclaude

Install

agr install @charlieviettq/interview-me --target claude

Writes 1 file into .claude/skills/, pinned to git-d1ca20bb.

  • .claude/skills/interview-me/SKILL.md

Document


name: interview-me description: "Extract user intent through structured one-question-at-a-time interviews before specs or implementation. Use when requirements are fuzzy, the user says "help me think through", or you need success criteria before planning." allowed-tools: Read, Glob, Grep

Interview Me

Structured intent extraction before specs, plans, or code. Complements clarify-underspecified with a stricter one-question cadence and explicit stop conditions.

When to use

  • User has a goal but not a spec
  • Multiple interpretations exist and batch questions would overwhelm
  • Before spec-driven-development or planning-and-task-breakdown

When not to use

  • User already provided a complete spec or ticket
  • Single factual question suffices
  • Emergency fix with clear repro steps

Workflow

  1. State purpose — one sentence on what you are extracting (goal, constraints, success criteria).
  2. Ask one question — highest-leverage unknown only; offer 2–4 concrete options when helpful.
  3. Listen and summarize — after each answer, one-line recap of what you learned.
  4. Stop when — you can write: objective, non-goals, constraints, acceptance criteria, and open risks.
  5. Hand off — produce a short intent brief or propose spec-driven-development.

Rules

  • Never ask more than one question per turn unless user explicitly requests a batch.
  • Prefer multiple-choice or examples over open-ended vagueness.
  • Do not start implementation until stop condition is met or user waives interview.
  • Flag assumptions explicitly; do not silently fill gaps.
  • If user jumps to implementation, pause and confirm intent or non-goals in one question.

Output template

## Intent brief
- **Objective:**
- **Non-goals:**
- **Constraints:**
- **Success criteria:**
- **Open risks:**
- **Next step:** spec / plan / spike / implement

Trustgrade A

  • passBody integrity

    Whether the stored document is plausibly the kind of file the artifact declares, rather than something fetched by mistake.

  • passType matchnot applicable to this artifact type

    Whether the artifact is really the kind of thing its metadata claims it is.

  • passFreshness

    How long since the source repository was last pushed to.

  • passPrompt injection

    Scans the artifact's own text for instructions aimed at your agent rather than at you.

  • passLicense

    Whether the source repository declares an SPDX license permissive enough to redistribute.

How the grade is calculated

Each check contributes 0 points when it passes, 1 when it warns, and 2 when it fails. The total maps to a letter:

  • Aevery check passed
  • Bone warning
  • Ctwo warnings
  • Dprompt injection or body integrity failed, or three warnings
  • Fone of those failed, and something else is wrong

These are automated hygiene checks, not a security audit, and not a dependency or vulnerability scan. A grade of A means nothing was flagged — not that the artifact is safe.

Versions

  • git-d1ca20bbc84f2026-07-31