mreq-define

Installation
SKILL.md

mreq-define

Run only when the user invoked this skill: typed /mreq-define (/mreq:mreq-define with the Claude Code plugin, $mreq-define in Codex), or asked for it by name. If you are reading this file on your own, stop and ask the user to run /mreq-define <request>.

Turn a request into one confirmed task document. Do not implement anything: mreq-apply does that, after the user chooses it. The one exception is "Without a document": the user chooses in step 4 to implement the restatement here, with no document.

Input: What the user wants: a feature, a bug, a refactor, or any other task. It can be vague.

Language

Write and reply in the language the agent instruction files (AGENTS.md, CLAUDE.md, ...) set. When they set none, use the language the user writes in. A reply that only picks an option (1, Accept the code) or gives a path does not change it. Use that language in every message, from the first line to the report, including the short notes between tool calls. Only fixed labels such as Open: stay as written.

Asking the user

When your agent has a tool that asks the user option questions, ask with it: AskUserQuestion in Claude Code, ask_user_question in Grok, ask_question in Antigravity, request_user_input in Codex. Ask up to 4 independent questions at once, or as many as the tool takes, the recommended option first and labelled "(Recommended)". When a question has more options than the tool takes, ask that question in plain text. When the tool returns an error (Codex allows it only in Plan mode), ask in plain text for the rest of this conversation. Without such a tool, ask in plain text as a numbered list, the recommended option first, and wait for the answer.

Installs
20
Repository
seanmars/mreq
GitHub Stars
1
First Seen
Nov 6, 2026