# skill-elicitation-agent

skill-elicitation-agent interviews you about the skill you have in mind, then hands back a full Claude skill spec: examples, triggers, file layout and test cases.

- subagent · `type=agent` · Claude Code, Codex, Cursor · ★ 1357 repo stars
- Author: qdhenry
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/qdhenry/Claude-Command-Suite/tree/HEAD/.claude/agents/skill-builder/skill-elicitation-agent.md
- Updated: 2026-03-01
- Security check: automated, 2026-11-16 (text) — no signs of malicious code found.
- HTML: https://affpapa.org/en/nevibecode/agents/skill-elicitation-agent

## What it does

The agent works in stages. First it asks for a couple of concrete cases: what a user would type and what should come out of it. Those cases get picked apart to see which code is rewritten every time (a candidate for `scripts/`), which docs are consulted again and again (`references/`), which boilerplate is reused (`assets/`). Questions about scope, complexity and tool access come after that.

What you get back is a spec document: metadata with a third-person description, the situations where the skill should fire on its own, a structure type (single `SKILL.md`, multi-file, tool-restricted, with assets), a plan for loading context in stages, and test cases. Before handing off, the agent checks the description, single responsibility, a minimal tool list and the examples.

It does not write the skill itself. The spec goes to `skill-generator-agent`, which builds the files, so without that agent you move the document into `SKILL.md` by hand. Tools on hand: `Read`, `Write`, `Grep`, `Glob` and `WebFetch`.

## Install

### Claude Code (native platform)

**Manually: put the files in place**
- File: `.claude/agents/skill-elicitation-agent.md` — For all projects use ~/.claude/agents/. Claude delegates by the description field or an @-mention

### Codex

**Manually: put the files in place**
- File: `.codex/agents/skill-elicitation-agent.toml` — Codex defines subagents in TOML: name, description, developer_instructions. Move the Markdown text into developer_instructions by hand. For all projects use ~/.codex/agents/

### Cursor

**Manually: put the files in place**
- File: `.cursor/agents/skill-elicitation-agent.md` — Cursor also reads .claude/agents/

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/qdhenry/Claude-Command-Suite/tree/HEAD/.claude/agents/skill-builder/skill-elicitation-agent.md

## When to use it
- turn a vague idea into a written skill spec
- collect concrete usage examples and trigger phrases
- decide between a single-file skill and one with scripts or assets
- check the skill description for third-person wording and trigger keywords
- plan staged context loading and test cases before generation

## Example prompts
- Draft a skill spec for database migrations, start with concrete usage examples
- Interview me about a code review checklist skill and write the specification
- Turn my PDF processing request into a skill spec with scripts and test cases

## Similar
- [requirements-elicitation](https://affpapa.org/en/nevibecode/skills/requirements-elicitation.md) — skill · ★ 587 repo stars
- [feature-forge](https://affpapa.org/en/nevibecode/skills/feature-forge.md) — skill · 4600 installs
- [ce-brainstorm](https://affpapa.org/en/nevibecode/skills/ce-brainstorm.md) — skill · 4200 installs
- [elicit](https://affpapa.org/en/nevibecode/skills/elicit.md) — skill · 2800 installs
- [prompt-optimizer](https://affpapa.org/en/nevibecode/skills/daymade-prompt-optimizer.md) — skill · 2200 installs
- [sparc-spec](https://affpapa.org/en/nevibecode/skills/sparc-spec.md) — skill · 733 installs

From the same set: [qdhenry/claude-command-suite](https://affpapa.org/en/nevibecode/by/qdhenry/claude-command-suite.md) — 112 cards

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md
