Review Brief logo

Review Brief

OrganizationPopular
every-app
review-brief

Produce a scannable review brief for a branch or PR so the maintainer can approve without reading the whole diff. Use when the user asks "how do I review this", "summarize the changes", "what should I look at", or wants to understand a PR in under two minutes.

Overview

Publisherevery-app
Repositoryopen-seo
Skill namereview-brief
Stars
22K
Forks
2.8K
Bundled files
Instructions only
LicenseMIT
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by every-app on GitHub. Read the source before you install it.

Installation

Install the Review Brief AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/every-app/open-seo/tree/main/.agents/skills/review-brief
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/every-app/open-seo.git /tmp/open-seo
mkdir -p .claude/skills
cp -r /tmp/open-seo/.agents/skills/review-brief .claude/skills/review-brief
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Review Brief in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Review Brief on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Review Brief is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Review brief

The reader will not read the diff. They will skim for 30 seconds, then read only the code that could hurt them. Optimize for that.

Process

  1. Get the diff (git log --oneline main..HEAD, git diff --stat main..HEAD) and read every non-trivial changed file in full.
  2. Find the invariant core: what was broken, what is true now. One sentence.
  3. Classify every hunk. Anything matching the crucial criteria below goes into the code section verbatim. Everything else gets one table row or is folded.
  4. Write the brief in the layout below. Every line number cited must be verified against the working tree.
  5. Run a second agent as a reviewer of the CODE, not the prose: "what is wrong, overstated, or missing?" Fix real bugs before delivering, and say you did.

Crucial code criteria

A hunk is crucial and must be shown as code if it touches any of:

  • Schema. Drizzle tables, columns, indexes, migrations, Zod schemas at a trust boundary, MCP tool input or output schemas.
  • Database writes. New or changed INSERT/UPDATE/DELETE, conditional WHERE clauses, anything that runs inside a workflow step or a cron.
  • Security. Auth checks, scope or org or project gating, secrets, external calls with credentials, anything that fails open.
  • Core logic. The branch or expression that decides the outcome the bug was about: a status assignment, a validation gate, a ranking comparator, a retry or timeout policy.
  • Jank. Anything you would be embarrassed to defend in review. String-matching a vendor error message to pick UI copy. A regex standing in for a parser. A hard-coded lookup table. A console.warn where a metric belongs. A swallowed catch. A magic timeout. A DB read added to work around a missing parameter. A fallback that quietly changes behaviour. Show it, name it as jank in the sentence above, and say whether it is acceptable for now or should block.

Show each as 5 to 20 lines, trimmed to the decision, with one sentence above saying what to notice and one line below naming the failure mode if it is wrong. Three to six excerpts is normal. If there are more than six, the PR should probably be split; say so.

Layout

One sentence stating what was broken and what is true now.

Before / after diagram. One ASCII flow, two columns, showing the user-visible path. Boxes are steps, arrows are what flows. Label the bug locations.

The code that matters. The crucial excerpts, in reading order by risk. This section is the point of the brief. Do not defer it to a follow-up.

Everything else, in one table. Sorted by risk:

| # | File | What it does | The one thing to check |

One row per remaining meaningful change. Fold registrations, renames, formatting, and fixtures into a single final row.

What does NOT exist. Bullets naming the things a reader would assume were added and were not: no migration, no new dependency, no change to a persisted shape, no new auth path.

Decisions you might disagree with. Two to four bullets. Each: the choice, the alternative, the cost of the choice. No defence longer than one sentence.

Status table. Whatever numbers gate the merge: tests, dry-run counts, "applied to prod: No".

Try it in two minutes. The two or three commands that exercise the change.

Branch state. Commits, pushed or not, PR link or "no PR yet".

Rules

  • No paragraph over two sentences anywhere.
  • Bold the first few words of every bullet so the list skims by its openers.
  • Tables for anything with three or more parallel facts.
  • Code goes in fences, never in prose. Name a file only when the reader has to go there.
  • Cut history, rejected alternatives, and migration narrative unless asked.
  • Never simplify into inaccuracy. A detail that changes what the reader would do stays, as one line.
  • Under 600 words of prose. Code excerpts and tables do not count.

Anti-patterns

  • A numbered walkthrough with paragraphs per stop and a "Check:" paragraph.
  • Describing a WHERE clause, an auth check, or a status expression in words instead of showing it.
  • Hiding jank in a "decisions you might disagree with" bullet. If it is in the code, show the code.
  • Saving the code for "want me to show you the code?" The reader asked for a review brief; the code is the brief.
  • A draft PR description inline. Put it in a file and link it.

Frequently asked questions

What does the Review Brief AI skill do?

Produce a scannable review brief for a branch or PR so the maintainer can approve without reading the whole diff. Use when the user asks "how do I review this", "summarize the changes", "what should I look at", or wants to understand a PR in under two minutes.

Why use Review Brief on TypingMind?

Because you install it once and use it with any model. Review Brief is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Review Brief in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/every-app/open-seo/tree/main/.agents/skills/review-brief. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Review Brief?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Review Brief?

As many as you like. As long as a model supports skills, you can use Review Brief with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Review Brief AI skill free?

Yes. It is published on GitHub by every-app under the MIT license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇