AGENTS.md examples from real projects

hashcatDeletes or force-pushesAGENTS.md

hashcat

hashcat/hashcat/AGENTS.md

code you edited. ## Never present output you did not run Every command, hash, exit code and line of output in a commit message, a pull request or an issue

angularReads credentialsAGENTS.md

angular-cli

angular/angular-cli/AGENTS.md

titles. ## Building - Make a local build of all packages: ```shell pnpm build --local ``` ## Coding Practices - **Imports:** - Always use the `node:` protocol for Node.js built-in imports (e.g., `node:fs`, `node

icsharpcodeAGENTS.md

ILSpy

icsharpcode/ILSpy/AGENTS.md

Debug Steps pane), so run it from a local build, not the installed tool. ## Code review - **A PR review and exploratory testing are different jobs.** A review is scoped

deepset-aiReads credentialsAGENTS.md

haystack

deepset-ai/haystack/AGENTS.md

tell them instead of opening another. In addition, there are rules inferred from previous code reviews. Follow them like a reviewer's note: they encode reasons, not ceremony, so when

No other project uses it yet

promptfooReads credentialsAGENTS.md

promptfoo

promptfoo/promptfoo/AGENTS.md

local secrets and the file exists. Review the output file for `success`, `score`, and `error` fields. With the default pass-rate threshold, exit code 0 means the eval

No other project uses it yet

MicrosoftAGENTS.md

winget-cli

microsoft/winget-cli/AGENTS.md

open a pull request or link a branch when code is available. ## Pull requests Before opening a pull request: 1. Review `CONTRIBUTING.md` and follow the repository PR template. 2. Keep

apacheAGENTS.md

flink

apache/flink/AGENTS.md

tests that do not exercise the change, padded commit messages) will be closed without review ## Code Review Guidelines When reviewing a PR or diff against this repo: - Look for opportunities

rollupAGENTS.md

rollup

rollup/rollup/AGENTS.md

manually write the pull request in their own words to show they care. ## Code Review Focus - **Ignore style/linting issues** in test sample files (`test/*/samples/`) except for `_config.js` files - Test

vaporAGENTS.md

vapor

vapor/vapor/AGENTS.md

tools (Copilot, Claude, Cursor, etc.) to help write or refine code is fine, but a human must understand, review, and take responsibility for every change in the PR. - **Fully automated

goauthentikAGENTS.md

authentik / website

goauthentik/authentik/website/AGENTS.md

tightly as possible for a true one-off: ` ` in Markdown/MDX (`// spellchecker:ignore ...` in code), or ` ` for a single line. Never disable the checker for a whole page. ## Deployment Deployment

No other project uses it yet

zulipAGENTS.md

zulip

zulip/zulip/AGENTS.md

open-source community, with maintainers who have consistently emphasized **high standards for codebase readability, code review, commit discipline, debuggability, automated testing, tooling, documentation, and all the other subtle details that

biomejsAGENTS.md

biome

biomejs/biome/AGENTS.md

After implementation, code generation, formatting, linting, and tests, review the complete change with [biome-code-review](./.agents/skills/biome-code-review/SKILL.md). - Use a fresh subagent when one is available. - Provide only the review scope

No other project uses it yet

every-appAGENTS.md

open-seo

every-app/open-seo/AGENTS.md

controls that keep reviewers from flagging intentional behavior. After a merge-ready or other code review verifies a finding, add to that file only when the finding exposes a recurring

kieranklaassenAGENTS.md

compound-engineering-plugin

kieranklaassen/compound-engineering-plugin/AGENTS.md

exception — they review the files, they don't re-read them for context. `ce-code-review`'s project-standards reviewer globs `CODING_STANDARDS.md`, the designated criteria source, and reads `CLAUDE.md`/`AGENTS.md

clockworklabsAGENTS.md

SpacetimeDB

clockworklabs/SpacetimeDB/AGENTS.md

code follows the style doc or is moving toward it. Read the doc in full before your first edit here, and follow it. - No marker: Core. A person reviews every

No other project uses it yet

activepiecesDeletes or force-pushesAGENTS.md

activepieces

activepieces/activepieces/AGENTS.md

surface needs the placeholder. ## Key Utilities (`@activepieces/shared`) `apId()`, `tryCatch()`, `tryCatchSync()`, `isNil()`, `spreadIfDefined()`, `spreadIfNotUndefined()`, `ActivepiecesError({ code, params })`, `SeekPage `, `formErrors`, `BaseModelSchema`, `chunk()`, `partition()`, `unique()`, `omit()`, `sanitizeObjectForPostgresql()` ## Testing ```bash npm run test-unit

No other project uses it yet

pascalorgAGENTS.md

editor

pascalorg/editor/AGENTS.md

pascal-3d` skill → `agent-surfaces.md` (same contract and knowledge as the hosted AI chat). - Reviewing a PR: `.agents/skills/review-architecture/SKILL.md`. - Opening a PR: `.agents/skills/open-pr/SKILL.md`. - Humans: `README.md`, `SETUP.md`, `CONTRIBUTING.md`. `CLAUDE.md`, `GEMINI.md` and `.github/copilot-instructions.md` symlink

neoclideAGENTS.md

coc.nvim

neoclide/coc.nvim/AGENTS.md

interface change. - Update doc/coc-api.txt after API change. - Update doc/coc-config.txt after coc.nvim configuration change. ## Code Review Rules - Report only consequential correctness, compatibility, lifecycle, concurrency, and performance issues. - Check that public coc.nvim

nextai-translatorReads credentialsAGENTS.md

nextai-translator

nextai-translator/nextai-translator/AGENTS.md

satisfy the pre-commit hook. ## Testing Guidelines Unit tests live next to the code (`__tests__/foo.test.ts` or `foo.spec.ts`) and run with Vitest. Mock remote APIs and keep snapshots deterministic, especially

highlightjsAGENTS.md

highlight.js

highlightjs/highlight.js/AGENTS.md

edited grammar code. ```js // good { match: /\bfoo\b/, scope: "keyword" } // avoid in new code { match: /\bfoo\b/, className: "keyword" } ``` ### Drop `relevance` when touching a mode `relevance` on modes is **deprecated