Codex skill and standard-library Python CLI for evaluating whether an open-source dependency is safe enough to adopt.
简体中文用户: 查看中文说明
English: Read the English guide
Open Source Adapter gives Codex a repeatable adoption workflow before a third-party library enters your codebase:
- evaluate repository health, license fit, governance, maintenance, and risk
- compare alternatives with Markdown, JSON, or CSV output
- run CVE checks, GitHub security posture audits, and minimal SBOM generation
- smoke-test a package in a temporary environment
- generate thin adapter-layer templates for common languages
Runtime code uses only the Python standard library.
python scripts\oss_cli.py eval https://github.com/axios/axios --json
python scripts\oss_cli.py score https://github.com/axios/axios --preset enterprise --json
python scripts\oss_cli.py compare https://github.com/axios/axios https://github.com/nodejs/undici --format csv --output compare.csv
python scripts\oss_cli.py security batch-osv PyPI requests| Command | Purpose |
|---|---|
eval |
Score a GitHub or GitLab repository across License, Activity, Community, Governance, and Risk |
score |
Re-score with custom weights or built-in default, enterprise, and startup presets |
compare |
Rank multiple repositories and export Markdown, JSON, or CSV |
analyze |
Extract README/CHANGELOG signals and assess maintenance status |
security |
Query OSV, audit GitHub security posture, or generate a minimal SBOM |
quickstart |
Install a package in a temp directory and run a smoke test |
integrate |
Generate adapter scaffolding for TypeScript, JavaScript, Python, Go, or Rust |
Direct script entrypoints remain available under scripts/, while scripts/oss_cli.py is the recommended unified interface.
SKILL.md routes user intent into focused prompt templates:
| Intent | File |
|---|---|
| Evaluate, score, compare | prompts/01-evaluate.md |
| Security, CVE, SBOM | prompts/02-security.md |
| README, CHANGELOG, maintenance | prompts/03-analyze.md |
| Quickstart, adapter generation | prompts/04-integrate.md |
| Maintenance, upgrade, rollback | prompts/05-maintain.md |
Examples live in examples/, reusable background material lives in references/, and regression tests live in tests/.
Machine-readable sample outputs:
examples/eval-requests.json: point-in-time output frompython scripts\oss_cli.py eval https://github.com/psf/requests --json.examples/compare-http-clients.csv: point-in-time output frompython scripts\oss_cli.py compare https://github.com/axios/axios https://github.com/nodejs/undici --format csv.
Scores and repository metadata can drift; rerun the commands before making an adoption decision.
python -m pytest tests/ -qCurrent baseline: 196 tests across Python 3.11, 3.12, and 3.14 in GitHub Actions.
Coverage includes scoring boundaries, platform adapters, CLI dispatch, compare JSON/CSV rendering, example fixture validation, cache behavior, README/CHANGELOG extraction, and security report regressions.
git clone https://github.com/ToYOhin/open-source-adapter.git
cd open-source-adapter
$destination = "$env:USERPROFILE\.codex\skills\open-source-adapter"
New-Item -ItemType Directory -Force $destination
Copy-Item -Recurse -Force .\* $destinationOn macOS or Linux:
cp -r open-source-adapter ~/.codex/skills/open-source-adapterCurrent release: v0.1.1.
See CHANGELOG.md, RELEASE_NOTES.md, SECURITY.md, and CONTRIBUTING.md.
MIT