oss-policy-kit 5.4.0

Open source starter kit for adopting OSS security baselines via policy packs, templates, evidence, and remediation.

OSS Security Policy as Code Starter Kit

Evaluate clone-visible OSS repository governance plus GitHub Actions, Azure Pipelines, and AWS CodeBuild/CodePipeline signals from a local clone. Platform truth outside the repository still requires evidence or manual review.

This project is intentionally small and explicit about trust boundaries. GitHub support is the most mature path today; Azure and AWS support are clone-based and evidence-driven rather than live platform verification.

Operational privacy: evaluation is local and clone-visible by default. API-backed evidence collection runs only when explicitly invoked, and should be scoped to the platform/repository being assessed.

Quick Links

• What's new in v5.4.0
• Current Release State
• Recommended First Command
• Quickstart
• CI/CD Integration
• Outputs
• Documentation
• CHANGELOG.md
• GitHub Releases
• Validation walkthrough
• CLI reference
• Results guide
• Bundled profiles overview
• Release hard-gate playbook
• v5.1.0 migration guide
• v5.0.0 migration guide

At A Glance

Area	What you get
Current release	v5.4.0 / Python package oss-policy-kit==5.4.0
Input	A local repository clone
Output	evaluation-report.json and evaluation-report.md (optional SARIF 2.1.0 via --sarif-output)
Core scope	Clone-visible governance and GitHub/Azure/AWS CI/CD signals
Profiles	github-*, azure-*, aws-* ladders (level-1..3, release-hardening-*) — list with python -m oss_policy_kit profiles; optional bundled hybrids github-aws-level-2 / github-azure-level-2 are advisory-only; not for use as release gates
Exceptions	Waiver registry with owner, reason, and expiry
Examples	Hardened and vulnerable sample repositories
Assurance model	Each control is labelled deterministic, signal, or evidence-backed; the value flows into reports/1.0 JSON and Markdown so consumers can reason about proof strength. See docs/profiles/overview.md.

Recommended First Command

After installing, run the hardened example first:

python -m oss_policy_kit evaluate --target ./examples/hardened-repo --profile github-level-1 --output-dir ./out/hardened --summary-only

This confirms the CLI, bundled profile data, example repository, and report generation path before you evaluate your own repository.

Two-line bootstrap (v5.4.0+)

For brand new adopters, init collapses the first run into two commands:

python -m oss_policy_kit init --target . --with-evidence --with-workflow
python -m oss_policy_kit evaluate --target .

init writes a persisted oss-policy-kit.yaml; evaluate reads the profile from it when --profile is omitted. See docs/cli-reference.md for all flags.

Use as a GitHub Action

The kit ships a composite GitHub Action so adopters can evaluate the bundled baseline on every pull request without managing Python in their build images:

- uses: lucashgrifoni/OSS-Security-Policy-as-Code-Starter-Kit@v5
  with:
    profile: github-level-1
    fail-on: fail

Inputs map 1:1 to CLI flags. Full reference and SARIF forwarding example in docs/github-action.md.

Current Release State

v5.4.0 is the current public release line. It is an additive minor release on top of v5.3.0: the JSON report contract remains reports/1.0, legacy 0.3 and 0.2 report shapes remain selectable, and the existing profile IDs stay stable. The release adds the init wizard, makes evaluate config-aware via oss-policy-kit.yaml, ships an official GitHub Marketplace action (action.yml), introduces the first native SAST integration (scan-sast + Semgrep adapter + SAST-SEMGREP-064 experimental control), and consolidates documentation. See CHANGELOG.md for full details.

Surface	Current state
Package	oss-policy-kit==5.4.0
GitHub Release	v5.4.0 is the current release target; v5.3.0, v5.2.0, and v5.1.0 remain available as predecessors
Default branch	master
License	Apache-2.0 (LICENSE + NOTICE)
Report contract	reports/1.0 by default; 0.3 and 0.2 remain selectable via --report-json-contract. 0.1 was removed in v5.0.0
Security workflow	Scanners run in Security CI/CD; SARIF upload is gated by ENABLE_CODE_SCANNING_UPLOAD=true so validation does not fail when Code Scanning upload APIs are unavailable

The repository is designed to be reproducible from a clean clone: install the package, run the built-in examples, and compare the generated JSON/Markdown reports.

What's new in v5.1.0

Highlights only — full detail in CHANGELOG.md and the additive migration notes in docs/v5.1.0-migration-guide.md. For the major v4 → v5 transition, see docs/v5.0.0-migration-guide.md.

• Audit-log streaming evidence. New experimental control AUDIT-STREAM-060 checks centralized audit log streaming evidence for hard-gate profiles, closing OWASP CICD-SEC-10 and improving AWS/Azure traceability alignment.
• Signed-provenance verification. New experimental control PROV-VERIFY-061 reads optional verification: evidence for signed build provenance and moves the SLSA Build L2 "Provenance signed" mapping to YES.
• Hard-gate coverage expanded without breaking contracts. Both new controls are wired into the 6 *-level-3 and *-release-hardening-3 profiles. reports/1.0, 0.3, and 0.2 remain stable.
• Evidence schemas are additive. New audit-log streaming and GitHub provenance schemas were added, while Azure/AWS provenance schemas only gained optional verification fields. Existing v5.0.0 evidence files continue to validate.
• Framework alignment refreshed. The docs now include SLSA v1.2 Source Track notes, an EU CRA mapping with an explicit no-certification honesty contract, and a NIST SP 800-218A AI out-of-scope note.
• Friendlier Scorecard input errors. Malformed --scorecard-json files now return exit 2 with a clear parse message instead of surfacing as Unexpected error.

There are no breaking changes in v5.1.0. The --fail-on semantics, JSON report contracts, Markdown report layout, and the surfaces of evaluate-many / recommend-profile / scaffold-evidence / collect-evidence / diff-reports are unchanged.

What This Kit Does

The starter kit helps you assess whether a repository shows a practical OSS baseline in files and workflows that are observable locally. It focuses on:

• SECURITY.md, CONTRIBUTING.md, LICENSE, CODEOWNERS, changelog presence
• clone-visible GitHub Actions, Azure Pipelines, and AWS build/pipeline structure and hygiene signals
• report generation for humans and pipelines
• policy lifecycle markers such as stable, experimental, and deprecated
• additive local evidence such as waivers and optional platform evidence files

It is not a universal application scanner, an OSPS certification engine, a compliance guarantee, or a substitute for threat modeling, secure code review, pentesting, or GitHub settings review. See docs/results-guide.md for the full applicability discussion.

Quickstart

Requires Python 3.12+.

Install (official channels)

1. From PyPI (recommended for most consumers):

   python -m pip install oss-policy-kit
   python -m oss_policy_kit --version
   

2. From GitHub Release artifacts — when wheel or sdist assets are attached to a release, download oss_policy_kit-*.whl or oss_policy_kit-*.tar.gz from the Releases page, then pip install /path/to/wheel. PyPI remains the primary install channel.

3. From source (contributors):

   python -m pip install -e ".[dev]"
   

Prefer python -m oss_policy_kit for CLI entry on Windows so you do not depend on Scripts\ being on PATH.

Run The Built-In Examples

python -m oss_policy_kit evaluate --target ./examples/vulnerable-repo --profile github-level-1 --output-dir ./out/vulnerable
python -m oss_policy_kit evaluate --target ./examples/hardened-repo --profile github-level-1 --output-dir ./out/hardened

What Success Usually Looks Like

• examples/hardened-repo on github-level-1: pass: 14 (all active controls in the profile)
• examples/vulnerable-repo on github-level-1: multiple fail
• github-release-hardening-1: typically pass plus one branch-protection result that remains manual-review-required or self-attested, depending on local evidence

For the full step-by-step demo (CLI help, profile discovery, fixture comparison, controls table, CI gating), see docs/validation-walkthrough.md.

CI/CD Integration

Starter workflows live under templates/workflows/:

• github-oss-policy-check.yml — baseline evaluate against github-level-1, fail the job when any control is fail.
• github-oss-policy-check-with-waivers.yml — same as above but passes --waivers ./waivers/waivers.yaml.
• github-oss-policy-check-level-2.yml — stricter github-level-2 profile.
• pipelines/azure/azure-pipelines.yml — Azure Pipelines example for lint, tests, package build, SBOM artifact generation, and self-check gating on a Linux/Ubuntu agent.

Typical CI command:

python -m oss_policy_kit evaluate --target . --profile github-level-1 --fail-on fail --output-dir ./oss-policy-reports

Exit codes in pipelines: 0 evaluation finished and the --fail-on threshold was not violated; 1 the gate tripped; 2 is usage or load errors. Keep the output directory as an artifact with if: always() so reviewers can inspect reports after a blocked run.

GitHub Release assets are the intended alternate install path when wheel or sdist artifacts are attached. PyPI remains the primary install channel, and the publish workflow generates both distribution artifacts and a CycloneDX SBOM as workflow artifacts for release evidence.

Outputs

Each evaluate run writes:

• evaluation-report.md — human-readable report (summary, controls table, per-control detail)
• evaluation-report.json — machine-readable contract (default reports/1.0)
• evaluation-report.sarif — SARIF 2.1.0, only when --sarif-output is passed

For the JSON contract details see docs/reports-contract-v1.0.md and the legacy docs/reports-contract-v0.3.md. For result statuses (pass, fail, manual-review-required, self-attested, etc.) and how to interpret each one see docs/results-guide.md.

Maintainer Self-Check

Run the kit against this repository:

python -m oss_policy_kit evaluate --target . --profile github-level-1 --output-dir ./out/selfcheck

Treat the generated self-check report as authoritative for the current revision.

Optional maintainer scripts under scripts/:

• consumer_smoke.py — install the wheel into a clean venv and exercise the CLI end-to-end.
• twine_check_dist.py — run twine check against dist/*.whl / dist/*.tar.gz.
• check_public_hygiene.py — fail-fast guard against private tokens leaking into public files.
• demo-video.ps1 — PowerShell helper to record a CLI demo GIF/video; not part of CI or packaging.

Documentation

• docs/README.md — documentation hub
• Project site — public product site
• docs/validation-walkthrough.md — full step-by-step demo with screenshots
• docs/cli-reference.md — full CLI reference
• docs/results-guide.md — how to interpret report statuses
• docs/profiles/overview.md — bundled profiles matrix, assurance vocabulary, decision tree
• docs/framework-alignment.md — cross-framework mapping (Scorecard, OSPS, OWASP CICD Top 10, SLSA v1.0, NIST SSDF, S2C2F, CIS SSCS, AWS Well-Architected, Azure DevOps Security)
• docs/profiles/github.md / docs/profiles/aws.md / docs/profiles/azure.md — operator guides by platform family
• docs/release-playbook-hardgate.md — release gate with real CLI commands
• docs/release-hardening-workflow.md — end-to-end L3 and release-hardening workflow with evidence paths
• docs/adoption-guide.md — baseline selection and expected outcomes
• docs/recommended-adoption-playbook.md — copy/paste adoption path
• docs/architecture.md — package boundaries, trust model, and evidence semantics
• docs/packaging-and-release.md — install, build, and distribution guidance
• docs/release-readiness.md — release gate and public launch checks

Contributing

See CONTRIBUTING.md.

Security

See SECURITY.md.

License

Apache-2.0. See LICENSE and NOTICE.