7.0 KiB
AGENTS.md
This file provides guidance to AI coding agents (Claude Code, etc.) when working with code in this repository. CLAUDE.md is a symlink to this file.
Core Principles (CRITICAL)
Respecting these principles is critical for every PR.
Less is more. The simplest solution is the best solution.
The action hierarchy for every change: Delete > Replace > Add. The best code change is a deletion. The second best is modifying what exists. Adding new code is the last resort.
- Minimal: The simplest solution that works. Do not over-engineer, over-abstract, or add code just in case. Three similar lines beat a premature abstraction. Avoid error handling for impossible states, feature flags, compatibility shims, or policy scaffolding unless they are truly required.
- Solve at the source: Do not hack fixes. Solve problems at their root. If something is broken, fix or remove the broken thing. Never patch over a broken abstraction, add workarounds, or add synchronization code for state that should not be duplicated.
- Delete ruthlessly: When replacing code, delete what it replaced. Remove unused imports, functions, types, files, and commented-out code. Git preserves history. Run the repo's relevant dead-code or cleanup check when available.
- Replace > Add: Modify existing code over adding new code. Edit existing files, extend existing components or functions with minimal parameters, and reuse existing utilities. If creating a new file, first prove it cannot fit cleanly in an existing file.
- Check existing: Search the entire repo before creating anything new. If a feature, component, helper, responder, workflow, or utility already solves a similar problem, reuse or adapt it and delete the duplicate path.
- Deduplicate: Do not duplicate existing code when updating the repo. Consolidate or refactor duplicates you find when it is in scope and low risk.
- Zero Regression: Do not break existing features or workflows unless the PR intentionally removes them with evidence.
- Production ready: All changes must be thoroughly debugged, validated, and production ready.
When fixing bugs, ask: "What can I delete?" before "What can I replace?" before "What should I add?"
PR Workflow
After opening a PR:
- Wait for the automated PR review and auto-format commit from Ultralytics Actions (
format.yml), then pull and address every finding. - Launch an independent adversarial review agent with cold context (just the PR diff and this file) to hunt for bugs, regressions, and Core Principles violations. Fix, push, and repeat with a fresh agent until one reports LGTM.
- Never fight other commits: Ultralytics Actions pushes auto-format and header commits, and multiple users may work on the same PR.
git pull --rebasebefore pushing; never force-push, reset, or revert commits you did not author. - After the PR merges, clean up: remove local worktrees and branches for it, then
git checkout main && git pull.
Commands
uv pip install -e ".[dev]" # install for development
pytest tests -v # run all tests
pytest tests/test_common_utils.py -v # run one test file
pytest tests/test_github_utils.py::test_name -v # run one test
pytest tests -v --cov=actions --cov-report=xml:coverage.xml # tests with coverage (CI command)
# Lint/format — mirrors the "Run Python" step in action.yml (source of truth if these drift)
ruff check --fix --unsafe-fixes --extend-select F,I,D,UP,RUF,FA --target-version py39 \
--ignore D100,D104,D203,D205,D212,D213,D401,D406,D407,D413,RUF001,RUF002,RUF012 .
ruff format --line-length 120 .
Notes:
- CI tests Python 3.8 and 3.14 on ubuntu and macos — code must stay 3.8-compatible. Use
from __future__ import annotationsfor modern type hints.
Architecture
This repo is two things at once:
- A Python package (
actions/) published asultralytics-actionson PyPI. Top-level modules (first_interaction.py,review_pr.py,summarize_pr.py,summarize_release.py,dependabot.py,github_report.py, etc.) are standalone scripts, most exposed asultralytics-actions-*CLI entry points inpyproject.toml[project.scripts]. - GitHub composite actions. The root
action.ymlis the main "Ultralytics Actions" marketplace action: it installs the Python package, then runs formatters (Ruff, Prettier, Biome, swift-format, dart format, codespell) and the CLI entry points conditioned on event type and inputs, then commits results back to the PR. Subdirectoriesretry/,cleanup-disk/,dependabot/,github-report/are standalone composite actions with their ownaction.yml+ README.
Key flow: GitHub workflow event → action.yml step (gated by github.event_name / github.event.action / inputs) → env vars (GITHUB_TOKEN, OPENAI_API_KEY, ANTHROPIC_API_KEY, MODEL, ...) → CLI entry point → module main()/run().
actions/utils/ is the shared core:
github_utils.py— theActionclass, the central abstraction. Initializes from GitHub Actions env vars (GITHUB_TOKEN,GITHUB_EVENT_NAME,GITHUB_EVENT_PATH), wraps REST (get/post/patch/...) and GraphQL requests with unified status checking, and provides high-level operations (PR diffs, labels, comments, discussions, alerts).openai_utils.py— AI provider abstraction supporting OpenAI and Anthropic. The provider/model is auto-detected from which API key env var is set; defaults live here as single source of truth (OPENAI_MODEL_DEFAULT,ANTHROPIC_MODEL_DEFAULT,PR_REVIEW_MODEL_DEFAULT,MODEL_COSTS). Also holds shared prompt-building and response sanitization.common_utils.py— URL/redirect checking, diff filtering, file-skip patterns, HTML comment removal.version_utils.py— PyPI/pub.dev version checks used for publish gating.
Most shared utilities are re-exported through actions/utils/__init__.py — keep __all__ updated when adding exports.
Self-hosting detail: .github/workflows/format.yml checks out the event's ref (PR head for pull requests, main for issue events) and runs uses: ./ — so PRs here dogfood both the Python package and action.yml itself before merge.
Conventions
- License headers (
# Ultralytics 🚀 AGPL-3.0 License - https://ultralytics.com/license) are added automatically by Ultralytics Actions (ultralytics-actions-headers, extensions inCOMMENT_MAP) — don't add or revert them manually. - Bump
__version__inactions/__init__.pywhen a PR changes package behavior — publishing to PyPI is gated on the version change (publish.yml). - Google-style docstrings, single-line summaries where possible; formatting is enforced by the repo's own action (
format.yml), which auto-commits fixes to PRs. - Tests use
unittest.mockto patch env vars and network calls, excepttests/test_urls.pywhich makes live HTTP requests. Modules listed in[tool.coverage.run] omitare excluded from coverage requirements. - Commits and PRs use plain git identity — no AI attribution, co-author lines, or generated-with footers.