AI Skill MartBrowse skillsPromptsProgramsLearnWhat is a Skill?
 
← Back to list

Agent Brief

A pre-delegation checklist that makes you write stakes, hypothesis, scope and a concrete definition of "verified" before handing work to an agent.

AutomationBeginner★ 537⑂ 41AI score 9/10Last updated: Nov 4, 2026

What it does

  • Forces a stakes call first: throwaway work gets handed off immediately; anything touching shipped code, users, money, permissions or data runs the full checklist.
  • Captures a one-line hypothesis before the agent starts, so a polished answer can't quietly become the default.
  • Has you brief in your own words: goal (outcome, not steps), who uses it, explicit out-of-scope, and constraints (versions, editable files, allowed commands).
  • Locks down what "verified" means before you see output — a named test passing, a command plus expected output, a log line or screenshot from the real app, a number recomputed from source data.
  • Makes you name the mode: is the agent thinking with you or for you? The second mode buys a stricter review.
  • Ships a paste-ready card for the top of any task.

Who it's for

  • Anyone who frequently spawns subagents or delegates whole PRs, analyses or docs.
  • Reviewers who struggle to tell "correct" from "plausible" in agent output.
  • Tech leads who want a shared delegation template for the team.

Examples

  1. Flaky test investigation: write "hypothesis: compaction.e2e is a race between the summary write and the next turn", define done as "that test passes and a new test fails on the old code", then delegate.
  2. Shipping PR: mark public API and lockfile as out of scope, define done as a command plus its expected output, then follow up with an evidence-first review.
  3. Analysis doc: set mode to "do it for me" and require the report order — commands run and their output first, then summary, then assumptions and decisions.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/MiniMax-AI/mcode-stack/HEAD/stacks/ronny/skills/agent-brief/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on AI Skill Mart.
Copy the stacks/ronny/skills/agent-brief folder from the GitHub repo MiniMax-AI/mcode-stack into my ~/.claude/skills/agent-brief/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/MiniMax-AI/mcode-stack.git /tmp/mcode-stack && mkdir -p ~/.claude/skills/agent-brief && cp -r /tmp/mcode-stack/stacks/ronny/skills/agent-brief/* ~/.claude/skills/agent-brief/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repo: git clone https://github.com/MiniMax-AI/mcode-stack.git /tmp/mcode-stack
  3. Create the skill folder: mkdir -p ~/.claude/skills/agent-brief
  4. Copy the files: cp -r /tmp/mcode-stack/stacks/ronny/skills/agent-brief/* ~/.claude/skills/agent-brief/
  5. Confirm ~/.claude/skills/agent-brief/SKILL.md exists.
  6. Restart Claude Code and try: "write the brief before I hand this task to a subagent".
  7. (Optional) Install the companion evidence-first-review skill from the same repo to cover the review step it points to.
View source on GitHub ↗License: MIT