create-pr
Skill1 starsChanged 7 months agoMITView on GitHub
Full end-to-end PR creation — branch setup, implementation, tests, commit, push, and GitHub PR
- Commits and pushes
196 other projects use a version of this.
Adopted by
Most recently taken up by
- Automattic/wp-calypsotheir own version, first seen Nov 16, 2026
- AtilaMedeiros/skillstheir own version, first seen Nov 16, 2026
- endalk200/skillstheir own version, first seen Nov 16, 2026
- hotcodepush-team/skillstheir own version, first seen Nov 16, 2026
- mlschodkowski/skillstheir own version, first seen Nov 16, 2026
- craigspaterson/agent-skillstheir own version, first seen Nov 15, 2026
Used with
Projects with it also use update-pr, act-on-feedback and 10 more.
Other files people use alongside this one.
Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
MicrosoftSkillupdate-pr
Update the pull request for the current session. Use when the user wants to push new changes to an existing PR.
Also in 9 projects that use this
MicrosoftSkillact-on-feedback
Act on user feedback attached to the current session. Use when the user submits feedback on the session's changes via the Submit Feedback button.
Also in 3 projects that use this
ModelEngine-GroupSkillcreate-release-note
从 PR/commit 中自动生成结构化的 Release Notes,并可选创建 GitHub Draft Release。支持 x.y.0 版本合并前一 minor 系列的 release notes。当用户要求生成发布说明、创建 release notes 时触发。参数为版本号和可选的上一版本号。
Also in 3 projects that use this
frostneySkillnative-nostalgia-stack
Applies the user's FreePascal toolchain and its build, formatting, hook, and test contracts while leaving project-specific mechanics local. Use when scaffolding or working in a FreePascal project that follows this toolchain.
Also in 3 projects that use this
frostneySkillsoftware-engineering-excellence
Applies the user's engineering standards during substantial technical work: preserve scope, use current evidence, delegate independent parts, and complete authorized outcomes. Use during substantial implementation, debugging, refactoring, or multi-part delivery work.
Also in 3 projects that use this
parkchuSkillstart-mission
넥스트스텝(NextStep) 미션을 처음 시작할 때, 미션 저장소를 fork·clone하고 최초 작업 브랜치(step1 등)까지 만들어 준다. gh CLI로 fork하고 clone까지 이어서 처리하며, gh가 없거나 인증이 안 돼 있으면 브라우저 Fork 안내로 폴백한다. 이미 fork/clone돼 있으면 재사용하고 이어서 진행한다(재실행 안전). 사용자가 "미션 시작하려는데 저장소 세팅해줘", "새 미션 시작할래", "fork부터 해줘", "/nextstep:start-mission" 등으로 요청할 때 사용한다.
Also in 3 projects that use this
- +1 more from microsoft
Other versions196 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 196
- microsoft/vscode★ 195ksame name
- vercel/next.js★ 144ksame name
- penpot/penpot★ 62ksame name
- remix-run/react-router★ 57ksame name
- moeru-ai/airi★ 51ksame name
- ToolJet/ToolJet★ 41ksame name
- ueberdosis/tiptap★ 39ksame name
- BabylonJS/Babylon.js★ 26ksame name
- kortix-ai/CompanyOS★ 21ksame name
- kortix-ai/suna★ 21ksame name
186 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- comment-prSkill
- lfx-proposalSkill
- lfx-research-architectSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
24 rules. 0 are common, 19 are distinctive, and 5 are very specific to this project.
0
Common
Most of the 179 edited versions we read keep these.
19
Distinctive
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 17
Follow every step below in order.
- Distinctive: line 28
If the working tree is dirty, stop and tell the user to stash or commit their changes first.
- Very specific: line 34
Look at the task description in $ARGUMENTS and derive a branch name that:
- Distinctive: line 55
Before writing a single line:
- Distinctive: line 57
Read every file that will be modified
- Distinctive: line 59
Look at existing comments in the files — match their tone, length, and placement exactly.
- Distinctive: line 66
Write the full implementation.
- Distinctive: line 68
No TODOs, no stubs, no placeholder logic
- Distinctive: line 72
No extra abstractions, no extra features beyond what was asked
- Distinctive: line 79
Write tests that cover the new behavior and any edge cases
- Distinctive: line 80
Run the test suite and make sure everything passes before moving on
- Distinctive: line 81
If tests fail, fix the code (not the tests) unless the tests themselves are wrong
- Very specific: line 87
Run whatever lint, type-check, or build commands the project uses.
- Distinctive: line 103
No mention of Claude, AI, or any tool
- Distinctive: line 104
No period at the end
- Very specific: line 119
Before running gh pr create, do a final check:
- Distinctive: line 125
If anything looks off, fix it and amend or add a follow-up commit before proceeding.
- Very specific: line 134
Follow the org convention: check recent merged PRs with gh pr list --state merged --limit 10 and match the format
- Very specific: line 135
If no clear convention, use: <type>: <short description> — e.g.
- Distinctive: line 139
No Claude or AI mention
- Distinctive: line 145
No em dashes (—) anywhere — use commas, colons, or just restructure the sentence
- Distinctive: line 146
No mention of Claude, AI, "generated", "implemented", "leveraged", "utilized", or any AI-adjacent phrasing
- Distinctive: line 150
Do not start with "This PR" or "This commit" — just describe what changed
- Distinctive: line 151
Do not use bullet points for everything — use prose where it reads naturally
“Common” is measured: kept by at least 6 in 10 of the 179 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Task
$ARGUMENTS
Instructions
Follow every step below in order. Do not skip any step. Do not proceed to the next step if the current one fails.
Step 1 — sync with master
git checkout master
git pull origin masterIf the working tree is dirty, stop and tell the user to stash or commit their changes first.
Step 2 — pick a branch name
Look at the task description in $ARGUMENTS and derive a branch name that:
- is kebab-case:
fix-login-redirect,add-export-csv,update-rate-limit-config - starts with a short type prefix:
fix/,feat/,chore/,refactor/,docs/ - is based on the actual work, not generic words
- contains no AI words: no "implement", no "enhance", no "leverage", no "utilize", no "streamline", no "optimize" unless that is literally what the task is
- is under 50 characters
- has no uppercase letters
Show the proposed branch name and ask the user to confirm before creating it.
Once confirmed:
git checkout -b <branch-name>Step 3 — understand the codebase before touching anything
Before writing a single line:
- Read every file that will be modified
- Grep for existing patterns: naming conventions, error handling style, import style, comment style
- Look at existing comments in the files — match their tone, length, and placement exactly. If comments in the repo are terse one-liners, write terse one-liners. If they are absent, write none.
- Understand the test setup: where tests live, how they are named, what testing library is used
Step 4 — implement the task
Write the full implementation. Rules:
- No TODOs, no stubs, no placeholder logic
- Handle every error path with context
- Match the exact code style of surrounding code — indentation, spacing, naming, everything
- Comments must match the existing comment style in the repo. Do not add comments just to explain what the code does. Only add a comment when the why is non-obvious.
- No extra abstractions, no extra features beyond what was asked
Step 5 — write or update tests
- Find where tests live for the code you changed
- Write tests that cover the new behavior and any edge cases
- Run the test suite and make sure everything passes before moving on
- If tests fail, fix the code (not the tests) unless the tests themselves are wrong
Step 6 — pre-commit checks
Run whatever lint, type-check, or build commands the project uses. Check package.json, Makefile, justfile, or CI config to find them. Fix any issues before committing.
Step 7 — commit
Stage only the files relevant to this task. Then commit with:
git commit -s -m "<one-line message>"Commit message rules:
- One line only — no body, no bullet points
- Imperative mood: "add", "fix", "update", "remove" — not "added", "fixed", "updates"
- Under 72 characters
- No mention of Claude, AI, or any tool
- No period at the end
- The
-sflag adds the DCO Signed-off-by line using the user's git config name and email — this is the only signoff, and it comes from the user's identity, not Claude
Step 8 — push
git push -u origin <branch-name>Step 9 — review everything before opening the PR
Before running gh pr create, do a final check:
git diff master..HEAD— confirm the diff is exactly what was intended, nothing extra- Confirm tests pass
- Confirm no debug logs, no commented-out code, no leftover TODOs
If anything looks off, fix it and amend or add a follow-up commit before proceeding.
Step 10 — write the PR
Open the PR with gh pr create. Follow every rule below exactly.
Title rules:
- Follow the org convention: check recent merged PRs with
gh pr list --state merged --limit 10and match the format - If no clear convention, use:
<type>: <short description>— e.g.fix: handle empty state in user list - Sentence case only — capitalize the first word and proper nouns, nothing else
- No trailing period
- No em dashes
- No Claude or AI mention
Description rules — read carefully:
- Written as a human writing to teammates, not as a bot summarizing a task
- Mild use of capitals: capitalize the first word of a sentence and proper nouns only
- No em dashes (—) anywhere — use commas, colons, or just restructure the sentence
- No mention of Claude, AI, "generated", "implemented", "leveraged", "utilized", or any AI-adjacent phrasing
- No signoff of any kind in the PR description body — no "Signed-off-by", no "Co-authored-by", no "Generated with" lines. The DCO signoff belongs only in the git commit (via
-s), never in the PR description. - Tone: direct, practical, like a developer who knows what they did and why
- Short paragraphs. No walls of text.
- Do not start with "This PR" or "This commit" — just describe what changed
- Do not use bullet points for everything — use prose where it reads naturally
Description structure:
Write it as plain prose. No fixed sections, no templates. Just describe what changed and why in a natural way, the same way you would explain it to a teammate over chat. If the approach needs a quick note, add it inline. Mention what was tested at the end in one or two sentences. That is all.
Final command:
gh pr create --title "<title>" --body "$(cat <<'EOF'
<description>
EOF
)"Step 11 — done
Print the PR URL and a one-sentence summary of what was done.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

