AGENT SKILL

explain-diff

Use when the user asks to explain, understand, or describe a code change: a branch, pull request, commit, diff, or patch, such as 'explain what branch X does' or 'what does this PR change', even when bro is also active or the diff is pasted inline. Also use it when the user did not understand a reply or a concept, or asks to explain it better or to give a real example; at the end of planned work, after its checks and review; and to write or update a pull request description.

License 许可
Apache-2.0
最近更新
2026-11-14
Stars
★ 1.6k
Forks
490

01 前言 Front Matter

name
explain-diff
description
Use when the user asks to explain, understand, or describe a code change: a branch, pull request, commit, diff, or patch, such as 'explain what branch X does' or 'what does this PR change', even when bro is also active or the diff is pasted inline. Also use it when the user did not understand a reply or a concept, or asks to explain it better or to give a real example; at the end of planned work, after its checks and review; and to write or update a pull request description.
allowed-tools
Glob, Grep, Read, Bash, Write, AskUserQuestion, Artifact, Skill
argument-hint
[--visual] [scope]
user-invocable
true

02 资源文件 2

03 SKILL.md 正文

Explain diff

People often no longer read the code that agents write. This explanation is how a person learns what was built, how it works, and why, well enough to own it. Write for someone who knows nothing about the project or the task: first invoke development-skills:bro for this output.

Gather the facts

Use the scope the user names: a branch, range, pull request, commit, or patch. Default to the work just finished. Read the request, the plan and chronicle, the final diff, the check results, and the review verdict. Check the behavior you describe in the code, and in real data and logs when they settle a claim. Call a claim you could not check unchecked, and say what is missing. Change no repository file.

Write the explanation

Write these parts in this order, and skip a part that has nothing to say:

  1. What changed: the result for the people or systems that use it compared with before, and the problem it solves, in two or three sentences.
  2. Concepts: each functional, business, or technical idea the solution rests on, such as an idempotency key or a dead-letter queue. Explain it in plain words, then say where this change uses it.
  3. Flow: follow one real case from input to result, with real values from the data, tests, or logs. Show what is read, stored, sent, or shown at each step.
  4. Edge cases and errors: what happens with missing or duplicate data, a timeout, or a failing dependency, and what a user or operator sees.
  5. Choices and limits: the alternatives the work rejected and why, the trade-offs, and what stays unchecked.

Explain each term the first time it appears, including common technical ones such as hash, lock, or transaction. Use the names a person sees or operates: endpoints, tables, queues, settings, screens, log messages. Leave out internal files, functions, classes, variables, and code settings unless the user asks for them. Draw a small text diagram when the flow branches. For planned work, keep it to about one screen. For a small change, write a few lines. At the end of planned work, give the explanation and the questions in their own message, before the next step such as align-docs.

Explain a reply or a concept

When the user did not understand a reply or a concept, explain that point only. Use the parts that fit and a real case from this project. Keep every fact of the earlier reply, and ask no understanding questions unless the user asks for them.

Check understanding

After planned work, ask two or three questions about the most important concepts: behavior, flow, or a technical choice. The explanation alone must be enough to answer them, without opening the code. Ask them in one call to the question tool. Each question is a concrete scenario with three plausible answers: one correct, and two that a reasonable person could believe. Shuffle the three answers and mark none as recommended. Add a fourth option last, "I don't know, explain", in the user's language. The free text field stays open for another answer.

  • Good: "A client sends the same order again after a timeout. How many orders are in the database?"
  • Bad: "Which function handles the retry?"

Recap the answers

After the user answers, write a recap in chat. For each question, show the question, the user's answer, and "correct", "wrong", or "skipped". "I don't know, explain" counts as skipped. For a wrong or skipped answer, explain the correct answer with a different example. A wrong answer can mean two things: the user lacks a concept, or the user wanted the system to behave that way. For each wrong answer, say which one the answer suggests, and why. When the user wanted a different behavior, check whether the code or the requirement is wrong before you correct the user.

Then ask with the question tool how to continue, and wait for the answer before the next step:

  • All answers correct: continue (recommended), more questions, or more explanation.
  • Some answers wrong: review the work on those points, more questions or explanation, or continue. Recommend the option that matches what the wrong answers suggest.

To review the work, settle the new requirement with the user, then change the work, repeat its checks and review, and explain the changed part again.

Visual mode

With --visual, or when the user asks for it, also build a page the user can step through: read visual mode. When the flow has states, ordering, or branches that are hard to follow in text, offer the page in one line.

Pull request description

Use the same parts for a reviewer who knows nothing about the project or the task: read pull request.

04 同仓库更多 Skills

hashgraph-online
hashgraph-online/awesome-codex-plugins

migrate-to-e2e

Moves an existing Playwright, Cypress, Selenium, Appium, Detox, Maestro, or TesterArmy suite to e2e one slice at a time, with the old suite running beside the new one until each ported test is proven. Keeps a MIGRATION.md progress file so any later session resumes where the last stopped. Optionally plants a defect to show both suites catch it. Use for /migrate-to-e2e, "migrate our tests to e2e", "port these Maestro flows", "move our TesterArmy tests", or when a project has an e2e config and an older suite still running.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

e2e

Writes, runs, and debugs end-to-end tests for web and mobile apps with e2e, the agentic test runner, which mixes agent.act goals with exact locator assertions on a Playwright browser or an iOS or Android device. Also drives a project's running app live over the e2e MCP server to look at a screen, check a UI change, take a screenshot or record a video, and try a locator; and runs bug bashes, parallel e2e explore runs with each bug proven by a repro test. Use when a project depends on e2e or has an e2e.config.ts, when asked for end-to-end, browser, mobile, or agentic UI tests, to check a change in the running app, to QA or bug bash an app or branch, or when an e2e run fails.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

create-verification-skill

Generate a project-local verify-<app> skill on top of e2e, so any coding agent can launch the app, check it is worth driving, exercise a feature the way a user does, keep evidence, and bug bash it. Writes the skill, a feature map with one explore charter per feature, and proves it once. Use for /create-verification-skill, "make a verification skill for this repo", or when a project has no scripted way to prove UI behavior.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

deja-search

deja-vu memory — search the user's past AI coding sessions with the deja CLI. Holds what they already settled and never wrote into the repo (naming rules, banned or already-tried tools, required checks before merge, preferred values, who owns what). Use before your first edit in a task, before you call a change done or ready to merge, before you recommend a tool, library, design or reviewer, before re-debugging an error that may already be solved, when they say 'didn't we fix this before', 'what did we decide about X' or 'I already have that', and before telling them something on this machine does not exist.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

deja-reflect

End-of-session reflection and memory curation with deja-vu. Review what was decided, what commands broke and what repaired them, and record durable project rules or decisions with deja remember so the next session starts with them.

★ 1.6k
hashgraph-online
hashgraph-online/awesome-codex-plugins

revid

Make, edit, export and post AI videos with Revid (revid.ai) through the Revid MCP server. Use whenever the user wants a short-form or long-form video made (from an idea, a script, a URL, an article, a PDF, a tweet, a product page, their own clips, or a song), wants to find what is going viral on TikTok, YouTube, Instagram or X before making something, wants to edit or re-export an existing Revid project, generate a standalone AI image or clip, make a series episode, or publish and schedule a finished video to connected social accounts. Covers setup (hosted MCP with OAuth, or an API key), the 36 tools, credits, polling and the rules for publishing. Routes to the more specific revid-* skills when one fits.

图像与多媒体Skill与智能体开发
★ 1.6k