Skip to main content
skill CX Codex

evidence-first-review — Claude Code skill: review and install

evidence-first-review is a Claude Code skill that audits an agent’s finished work from command logs and test evidence. It runs before a PR merge or an analysis is accepted.

MIT ★ 741 updated 2026-11-11

Install into Claude Code · Plugin marketplace All methods and platforms ↓
/plugin marketplace add MiniMax-AI/mcode-stack
/plugin install ronny@mcode-stack
Open on GitHub

What evidence-first-review does

evidence-first-review audits an agent’s completed task by examining the transcript or tool log before the final summary. It checks whether commands ran against the final code, then states what each passing check establishes and what remains outside its reach.

The skill also surfaces silent choices, asks which scope was left out, and tries to falsify the result. For analysis or documentation, it calls for opening two cited sources and recomputing one number behind the conclusion. A clean-context diff review is a separate job, covered by pairing this skill with independent-review.

What it helps with

  • verify that tests ran after the final edit
  • map each green check to the property it actually proves
  • surface defaults and scope decisions made by the agent
  • find missing evidence before merging a PR
  • recompute a source-based number behind an analysis

Example prompts for the agent

  • Review the agent’s finished work from the transcript and state what each command proves.
  • Read the tool log before merging the PR and look for a check that skipped the relevant suite.
  • List the agent’s silent decisions and try to falsify the result before accepting its report.

How to install evidence-first-review

1Plugin marketplace

/plugin marketplace add MiniMax-AI/mcode-stack
/plugin install ronny@mcode-stack

This installs the whole ronny plugin with all its skills

2Skills CLI (Vercel)

npx skills add MiniMax-AI/mcode-stack --skill evidence-first-review -a claude-code

3Manually: put the files in place

~/.claude/skills/evidence-first-review/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/evidence-first-review/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about evidence-first-review

How do I install evidence-first-review in Claude Code?

Add the marketplace with /plugin marketplace add MiniMax-AI/mcode-stack, then install the skill with /plugin install ronny@mcode-stack.

What does evidence-first-review do?

It reads the transcript or tool log before the agent’s summary, links each check to the claim it supports, and identifies what the check cannot establish. It also reviews silent decisions and tests whether the result could be wrong.

How is evidence-first-review different from independent-review?

evidence-first-review examines evidence from an agent’s completed work and its summary. independent-review is the companion for reviewing the diff itself in a clean context.

Source text

SKILL.md · MIT
---
name: evidence-first-review
description: Review an agent's finished work by reading the evidence before its summary, naming what each check proves and what it misses, pulling out the decisions it made silently, and trying to falsify the result before accepting it. Use when a MiniMax Code session, subagent or cloud agent reports a task done, before merging its PR or acting on its analysis.
---

# Evidence-First Review

"All tests pass, ready to merge" is a claim, not evidence. We read what the
agent actually ran before we read what it says about it. The bar for accepting
the work: we could defend it to the person who will maintain it, and we can
say what our checks would miss.

This skill is about owning the result. For a clean-context code review of the
diff itself, pair it with `independent-review`.

## 1. Evidence before summary

Open the transcript or tool log, not the final message. For each command that
supports "done": did it run on the final code, did it exit 0, does the output
say what the summary says? Typical gaps we find:

- tests run before the last edit;
- a filter like `-t compaction` that skipped the suite that matters;
- "fixed" with no reproduction before the fix;
- a build that passed only because a step was skipped.

## 2. For each green check, one property it proves and one it does not

| Check | Proves | Does not prove |
| --- | --- | --- |
| New unit test passes | the function returns X for this input | the caller passes that input; the test was written with the same wrong assumption |
| `pnpm typecheck` | types line up | runtime behaviour |
| E2E with a mock model | the loop wires up | anything about real model output |

If we cannot fill the right column, we do not understand the check yet.

## 3. Ask for the decisions list

Ask the agent: which defaults did you pick, what did you leave out of scope,
and where did you reinterpret the requirement because two parts clashed? These
are the choices that end up in production without anyone deciding them. Each one
is accepted, changed or sent back.

## 4. Try to break it

Ask: what would have to be true for this to be wrong? Then check that one thing
ourselves. For analysis and docs, open two of the cited sources and recompute
one number the conclusion depends on.

## 5. Make sure we understand it

- Pick a few lines of the diff at random and explain them without help.
- For a large change, ask the agent for a short report that ends in three
  questions about the change; we answer them before merging.
- For a contested call, have a fresh subagent argue the other side, then
  decide ourselves.
- Last question to a fresh context: "What did I miss?"

## 6. Decide

- **Accept**: the evidence matches, decisions are reviewed, we can name the gap.
- **Ask for more**: missing evidence or an unexplained decision.
- **Get a human review**: high stakes and we cannot say what the checks miss.

## Output

When we report the work to others, we lead with the evidence (commands and
results), then: what was checked, one thing that was not, and the key
decisions.

## References

- Addy Osmani, ["Critical thinking when agents do the work"](https://x.com/addyosmani/status/2117512471930167377)

© Ronny (MiniMax Code team) · MIT · MiniMax-AI/mcode-stack

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set27

The whole set →
trajectory-review trajectory-review is a Claude Code skill that examines an agent’s path to a result. It runs after a task or eval finishes, fails, or changes outcome. CX Codex +1 ★ 741repo stars harness-experiment Harness-experiment is a Claude Code skill that runs controlled experiments to validate agent scaffolding changes. It triggers when adjusting prompts, tools, or subagent layouts for performance tracking. CX Codex +1 ★ 741repo stars independent-review independent-review is a Claude Code skill that reviews changes from a clean context before merge. It compares the diff with the requirement and test output. CX Codex +1 ★ 741repo stars agent-brief The agent-brief skill for Claude Code structures task delegation to AI agents. It forces clarity on stakes, hypotheses, and verification criteria before an agent starts working. CX Codex +1 ★ 741repo stars pr-description pr-description is a Claude Code skill that drafts and verifies pull request descriptions for GitHub and GitLab. It runs when opening a PR or when MCode Review reports missing sections. CX Codex +1 ★ 741repo stars subagent-split-check subagent-split-check is a Claude Code skill that decides whether work belongs in a subagent or the main loop. It checks context boundaries before a split. CX Codex +1 ★ 741repo stars context-tax-audit The context-tax-audit skill for Claude Code inventories injected data in agent requests to reduce token usage. It identifies redundant prompt blocks and configurations that hinder prompt cache performance. CX Codex +1 ★ 741repo stars long-session-check long-session-check is a Claude Code skill that audits long-session context and memory handling. It starts when constraints get lost or compaction has happened several times. CX Codex +1 ★ 741repo stars

Project sponsors

Companies that keep this analytics open