
create-pr
Create a pull request with Conventional Commits formatting, a templated body, and local verification. Use when the user asks to create a PR, open a PR, submit changes for review, or put code up for review.
- Installs
- 109
- Last 24h
- —
- Trending rank
- —
- Security
- Not audited
Install
Works with every agent the skills CLI supports.
npx skills add https://github.com/factory-ai/factory-plugins --skill create-pr- 01
Install
Run the command in your project. The CLI asks which agents to install for.
- 02
Describe the task
Ask your agent for something this skill covers. It loads create-pr automatically when relevant.
- 03
Review the result
The agent follows the SKILL.md steps, scripts and references bundled with the skill.
SKILL.md
The instructions your agent reads when the skill is triggered.
Create Pull Request
Create a PR with proper conventions: local verification, Conventional Commits title, a templated body, and an optional linked ticket. This skill is language- and framework-agnostic — substitute your project's actual build, lint, test, and format commands where examples are shown.
Prerequisites
Before starting, verify:
- Current branch has commits not on the base branch (
git log origin/<base-branch>..HEAD --oneline) - Branch is pushed to remote (
git push -u origin HEADif not) - No uncommitted changes that should be included (
git status)
Workflow
1. Understand the Changes
Run in parallel:
git log origin/<base-branch>..HEAD --oneline
git diff origin/<base-branch>..HEAD --stat
Determine:
- What changed: Which modules, packages, services, or directories were modified
- Change type:
feat,fix,docs,refactor,test,chore,perf,ci,build,revert - Scope: Primary module/package/service affected (use directory name or
monorepo/repofor cross-cutting changes) - Is this a code change?: If the PR modifies source code (not only docs, markdown, or config-only changes), run the local verification checklist in step 2 before creating the PR.
2. Local Verification (for code changes)
Skip this step if the PR only touches documentation, markdown files, or other non-code files. For any change that touches source files, run your project's verification commands locally before creating the PR.
Discover the commands by reading the repo root (e.g. Makefile, package.json, pyproject.toml, Cargo.toml, go.mod, build.gradle, mix.exs, Gemfile, composer.json, justfile, Taskfile.yml, README.md, or the CI workflow config). Use filter/target flags where available (e.g. turbo --filter, nx --projects, pnpm --filter, bazel //path/..., cargo -p <crate>, pytest <path>, go test ./<pkg>/...) to run only the affected portions — it is faster than running the whole repo.
Common verification categories to run when applicable:
Typecheck / Compile
Run the project's static type check or compile step if it has one.
Examples across ecosystems (use whatever the repo defines):
- TypeScript:
npm run typecheck,pnpm -r typecheck,tsc --noEmit - Python (typed):
mypy .,pyright,ty check - Rust:
cargo check - Go:
go build ./...,go vet ./... - Java/Kotlin:
./gradlew compileJava,./mvnw compile
Lint / Format
Run the project's linter and formatter. Prefer an autofix target if one exists.
Examples:
- JS/TS:
npm run fix,npm run lint,eslint .,prettier --check . - Python:
ruff check --fix .,ruff format .,black .,flake8 - Rust:
cargo clippy --all-targets,cargo fmt --check - Go:
golangci-lint run,gofmt -l . - Shell:
shellcheck,shfmt -d .
Tests
Run the unit/integration tests for affected packages.
Examples:
- JS/TS:
npm run test -- --filter=<workspace>,pnpm -r test,vitest run <path>,jest <path> - Python:
pytest <path>,tox -e <env>,python -m unittest - Rust:
cargo test -p <crate> - Go:
go test ./<pkg>/... - Java/Kotlin:
./gradlew test,./mvnw test - Ruby:
bundle exec rspec <path>,rake test
Additional checks (run when relevant)
- Unused exports / dead code: Run your project's dead-code check if it has one (e.g.
knip,ts-prune,vulturefor Python,deadcode/unusedfor Go,cargo udepsfor Rust). - Dependency hygiene: Run your project's dependency check if it has one (e.g.
depcheck,pip check,cargo audit,bundle audit). - Lockfile in sync: If you modified any dependency manifest (
package.json,requirements.txt,pyproject.toml,Cargo.toml,go.mod,Gemfile, etc.), run the install command (npm install,pnpm install,uv sync,poetry lock --no-update,cargo update -w,go mod tidy,bundle install) and commit any lockfile changes. CI commonly fails if the lockfile is out of date. - Generated code / codegen: If the repo has an OpenAPI spec, protobuf, GraphQL schema, SQL migrations, or any other generated artifacts, regenerate and commit any changes.
- Style / asset linters: Run stylesheet linters (
stylelint, etc.) or asset linters if you changed those files. - Security scans: Run any security/secret scanners configured in the repo (
trivy,semgrep,gitleaks, etc.).
3. Link to a Ticket (optional)
If your org uses an issue tracker, ask the user whether to:
- Create a new ticket: Use the appropriate tool (Linear, Jira, GitHub Issues, etc.)
- Link an existing ticket: Ask for the identifier (e.g.
TEAM-1234,JIRA-567,#42) - Skip: Only if user explicitly says no ticket is needed
Most CI systems can be configured to require the ticket identifier in the PR body. Follow your org's convention.
4. Format PR Title
Follow Conventional Commits: type(scope): description
type:feat,fix,docs,refactor,test,chore,perf,ci,build,revertscope: Name of the affected module/package/service/directory, ormonorepo/repofor cross-cutting changes. Multiple scopes can be comma-separated:fix(a, b, c): ...
Examples:
feat(web): add dark mode togglefix(cli, daemon): load shell env at entrypointfix(api): handle nil response from upstreamchore(repo): bump dependencies
5. Generate PR Body
Fill in all sections from your PR template. A typical template has four sections:
## Description
<concise summary of what changed and why>
## Related Issue
Closes TEAM-XXXX
<!-- or: Part of TEAM-XXXX -->
## Potential Risk & Impact
<list risks, performance implications, technical debt>
<!-- Use "N/A" only if truly no risk -->
## How Has This Been Tested?
<describe testing performed: unit tests, manual testing, typecheck, lint>
6. Create the PR
gh pr create \
--base <base-branch> \
--head <branch-name> \
--title "<type>(<scope>): <description>" \
--body "<generated body>"
If the body is long, write it to a temp file and use --body-file:
gh pr create --base <base-branch> --head <branch> --title "..." --body-file /tmp/pr-body.md
7. Report Result
Return the PR URL to the user.
CI Checks Reference (template)
These are typical check categories that run on every PR. Map them to your repo's actual commands when adapting this skill.
Always-run checks
| Category | What it does | How to find the local command |
|---|---|---|
| Typecheck / compile | Verifies the project compiles or passes static types | Check package.json, Makefile, pyproject.toml, Cargo.toml, go.mod, CI config |
| Lint | Enforces code style / correctness rules | Check for lint, check, or equivalent scripts in the repo root |
| Format | Enforces consistent formatting | Check for format, fmt, prettier, black, gofmt, rustfmt, etc. |
| Tests | Runs unit and integration tests | Check for test script / target |
| Dead code / unused exports | Flags unused code | Check for knip, ts-prune, vulture, cargo udeps, etc. |
| Dependency check | Flags unused / vulnerable dependencies | Check for depcheck, audit, cargo audit, etc. |
| Lockfile in sync | Fails if lockfile is stale relative to the manifest | Run your package manager's install command and commit the lockfile |
| PR Conventions | Validates branch name, semantic title, ticket presence | Follow the formatting rules above |
Conditional checks (run only when affected files change)
- API / schema validation: Triggered by API or schema changes. Regenerate locally.
- Platform-specific builds: Triggered when desktop/mobile/embedded targets are affected.
- E2E tests: Triggered when the consumer app or top-level binary is affected.
Typical PR conventions CI enforces
- Branch name: Max length, allowed characters (e.g.
[A-Za-z0-9/-]). - Title: Conventional Commits format with a valid scope.
- Ticket reference: PR body must contain a ticket identifier (often skipped for
chore:andrevert:types).
Common Mistakes to Avoid
- Wrong base branch: Use the branch your org takes PRs into (e.g.
dev,main,develop,trunk). - Missing scope: PR title CI check often requires a valid scope.
- Missing ticket reference: Description must reference your ticket ID for CI to pass (except
chore:/revert:). - Forgetting to push: Branch must be on remote before
gh pr create. - Lockfile drift: Always run the install command and commit lockfile changes after dependency changes.
- Skipping local checks on code PRs: Typecheck/compile, lint, and tests should be run locally before sending out code changes to catch issues early and avoid CI round-trips.
- Uncommitted generated artifacts: After API/schema changes, regenerate and commit.
FAQ
What does the create-pr skill do?
Create a pull request with Conventional Commits formatting, a templated body, and local verification. Use when the user asks to create a PR, open a PR, submit changes for review, or put code up for review.
How do I install create-pr?
Run npx skills add https://github.com/factory-ai/factory-plugins --skill create-pr in your project. The skills CLI detects the agents you use and copies the skill into the right folder. For Claude Code you can also place the skill folder at ~/.claude/skills/create-pr/ (personal) or .claude/skills/create-pr/ (project).
Which AI agents can use create-pr?
Any agent that supports the open Agent Skills format (a folder with a SKILL.md file), including Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, Windsurf, Cline, OpenCode and more.
Is create-pr safe to install?
No partner security audit is available yet. Skills can run commands and scripts, so read the SKILL.md and any bundled files before installing.
Who maintains create-pr?
It is published by factory-ai in factory-ai/factory-plugins, and it is an official skill from the maker of the technology it covers. It has 109 recorded installs.
More from factory-ai/factory-plugins
no-use-effect
factory-ai/factory-plugins
Enforce the no-useEffect rule when writing or reviewing React code. ACTIVATE when writing React components, refactoring existing useEffect calls, reviewing PRs with useEffect, or when an agent adds useEffect "just in case." Provides the five replacement patterns and the useMountEffect escape hatch.
human-writing
factory-ai/factory-plugins
Remove signs of AI-generated writing from text to make it sound more natural and human-written. Use when editing or reviewing any form of document including: markdown, technical docs, emails, blog posts, PRDs, or any dedicated writing content. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases.
simplify
factory-ai/factory-plugins
Open the skill to read its SKILL.md instructions.
frontend-design
factory-ai/factory-plugins
Open the skill to read its SKILL.md instructions.
visual-design
factory-ai/factory-plugins
Open the skill to read its SKILL.md instructions.
security-review
factory-ai/factory-plugins
Open the skill to read its SKILL.md instructions.
Popular code quality & workflow skills
grill-me
mattpocock/skills
A relentless interview to sharpen a plan or design.
improve-codebase-architecture
mattpocock/skills
Scan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.
handoff
mattpocock/skills
Compact the current conversation into a handoff document for another agent to pick up.
domain-modeling
mattpocock/skills
Build and sharpen a project's domain model. Use when discussing codebase terminology, writing or editing a GLOSSARY.md, or recording or editing an ADR.
codebase-design
mattpocock/skills
Shared vocabulary for designing deep modules. Use when the user wants to design or improve a module's interface, find deepening opportunities, decide where a seam goes, make code more testable or AI-navigable, or when another skill needs the deep-module vocabulary.
diagnosing-bugs
mattpocock/skills
Diagnosis loop for hard bugs and performance regressions. Use when the user says "diagnose"/"debug this", or reports something broken/throwing/failing/slow.