ds-pr — Claude Code skill: what it does, how to install
ds-pr — a skill for Claude Code and other platforms that files a concise pull request for the current branch when asked to open a PR.
MIT 534 installs ★ 320 updated 2026-10-28
npx skills add aicodeking/deepseek-stack --skill ds-pr -a claude-codeWhat ds-pr does
The agent checks for existing pull requests, diffs the branch against a base such as origin/main, and follows recent commit formats. The title explains the user benefit, while the description covers the problem, fix, and proof. Before publishing, the agent strips secrets, tokens, and personal data, then prints the resulting GitHub URL.
Operations rely on gh pr list and gh pr create, falling back to the REST endpoint gh api repos/<owner>/<repo>/pulls on GraphQL rate limits. The tool requires the GitHub CLI installed in the environment.
The skill triggers whenever the user asks to open, file, or create a PR, or runs /ds-pr.
What it helps with
- file a concise pull request for a branch
- check for existing pull requests on a branch
- strip secrets and tokens from the description
- fallback to REST API on rate limits
Example prompts for the agent
- file a pull request for this branch
- /ds-pr
- open a PR for my changes
How to install ds-pr
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-pr -a claude-code2Manually: put the files in place
~/.claude/skills/ds-pr/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/ds-pr/
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-pr -a codex2Manually: put the files in place
.agents/skills/ds-pr/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-pr -a cursor2Manually: put the files in place
.cursor/skills/ds-pr/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install aicodeking/deepseek-stack skills/ds-pr2Manually: put the files in place
.github/skills/ds-pr/SKILL.md1Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-pr -a gemini-cli2Platform CLI
gemini skills install https://github.com/aicodeking/deepseek-stack --path skills/ds-prManually: put the files in place
.devin/skills/ds-pr/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add aicodeking/deepseek-stack --skill ds-pr -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about ds-pr
How to install ds-pr in Claude Code?
Run npx skills add aicodeking/deepseek-stack --skill ds-pr -a claude-code, then reference ~/.claude/skills/ds-pr/SKILL.md.
What does ds-pr do?
The skill generates and submits a clean pull request for the current branch, validates the diff, and removes sensitive data.
What is needed to run ds-pr?
The GitHub CLI (gh) tool must be installed and authenticated.
Source text
SKILL.md · MIT
--- name: ds-pr description: "Files a short, clear pull request for the current branch. Use when the user asks to open, file or create a PR, or says /ds-pr." --- Open one real PR whose title and description a busy reviewer can read in ten seconds. ## Steps 1. Check if a PR already exists for this branch (`gh pr list --head <branch>`). If it does, update it. 2. Diff the branch against its base, usually `origin/main`. Confirm the diff matches the goal and holds nothing unrelated. 3. Read the titles of a few recently merged PRs and copy their format. 4. Write a title that says what changed for the user and why it matters, not which function moved. Example: "Cut websocket frames by 70% with compression" beats "Negotiate permessage-deflate". 5. Write the description in this order: - The problem, in one or two plain sentences taken from the user's request. - The fix, in one or two sentences. - One line on how it was proven, if you have proof. - A closing line: `Created with <model> in <harness>.` Use the model name the harness reports. If it reports none, name only the harness. 6. Scan the title, body and any attachments for secrets, tokens, private paths and personal data. Remove them. 7. Create the PR as ready for review, not a draft, so review bots run. 8. Print the PR URL that GitHub returned. ## Stop when - GitHub returned a PR URL. Hand off to `ds-watch-pr` if the user asked for watching. - `gh pr create` fails twice. If the error is a GraphQL rate limit, retry once through REST: `gh api repos/<owner>/<repo>/pulls -f title=... -f head=... -f base=... -F [email protected]`. If that fails too, report the error. ## Rules - No file-by-file change lists in the description. - Do not post follow-up comments unless the request needs one. Use `ds-pr-comments` when it does. - Never guess a model name.
© aicodeking · MIT · aicodeking/deepseek-stack