oss-policy-kit 6.2.0

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

OSS Security Policy as Code Starter Kit

Pass/fail security policy gates for OSS repositories, with explicit assurance grading and framework mappings.

At a Glance

oss-policy-kit evaluates a local repository clone plus optional evidence files, then emits Markdown, JSON, and optional SARIF reports for humans and CI gates.

Current release	Bundled profiles	Controls	CLI commands	Python
v6.0.0	52	212	17	3.12+

Use it when you need a local-first gate that combines repository governance, CI/CD hardening, release posture, scanner evidence, waivers, and framework-oriented reporting. It is not a vulnerability scanner, certification engine, or legal compliance guarantee.

Quickstart

python -m pip install oss-policy-kit
python -m oss_policy_kit init --target . --with-evidence --with-workflow
python -m oss_policy_kit evaluate --target . --profile github-level-1 --fail-on fail

The evaluation writes:

• evaluation-report.md for review.
• evaluation-report.json for automation.
• evaluation-report.sarif when --sarif-output is set.

First-time tutorial: docs/tutorial-first-pr-gate.md. Compact CLI reference: docs/quickstart-15-min.md.

What It Does

• Evaluates bundled policy profiles against a repository clone.
• Uses optional evidence under .oss-policy-kit/evidence/ for platform-only facts.
• Composes signals from local files, workflows, SARIF/JSON scanner outputs, waivers, and release evidence.
• Labels controls by assurance type: deterministic, signal, or evidence-backed.
• Supports Markdown, JSON report contracts, and optional SARIF for code-scanning workflows.
• Keeps waivers visible with owner, reason, and expiry metadata.

What It Does Not Do

• It does not certify CRA, SLSA, OSPS, SSDF, or AI Act compliance.
• It does not replace SAST, SCA, secrets scanning, threat modeling, secure code review, pentesting, or live platform review.
• It does not prove branch protection, rulesets, MFA, cloud posture, or registry settings unless you provide API-backed evidence.
• It does not claim SLSA Build L3. The current trust model is documented in docs/supply-chain-verification.md.

Core Capabilities

Area	Included
Repository governance	LICENSE, SECURITY, CONTRIBUTING, CODEOWNERS, branch protection evidence, release hygiene
CI/CD posture	GitHub Actions, Azure Pipelines, AWS CodeBuild/CodePipeline, GitLab CI signals
Release hardening	OIDC publishing, provenance evidence, artifact verification, source-built container flow
Scanner composition	SARIF/JSON ingestion for tools such as zizmor, OSV-Scanner, Gitleaks, Scorecard, and Semgrep
Framework mapping	OSPS, NIST SSDF, SLSA, S2C2F, OWASP CI/CD, EU CRA, EU AI Act readiness signals
AI and agent security	AI agent source-side checks, MCP server security, OWASP Agentic ASI mapping
Exception handling	Waiver registry with reason, owner, scope, and expiry

Profiles

List bundled profiles:

python -m oss_policy_kit profiles

Common starting points:

Profile	Use when
github-level-1	First GitHub repository gate
github-level-2	Stricter GitHub governance and CI/CD posture
oss-publish-readiness-1	Release/publish readiness for OSS packages
appsec-sast-sca-1	Compose SAST/SCA/secrets scanner evidence
osps-baseline-2026-1	OpenSSF OSPS Baseline 2026-oriented review
cra-eu-ready-2-1	EU CRA Article 13/14 readiness signals
ai-agent-baseline-1	Source-side checks for AI agent repositories
appsec-mcp-server-1	MCP server security readiness

Full profile guide: docs/profiles/overview.md.

GitHub Action

- uses: lucashgrifoni/OSS-Security-Policy-as-Code-Starter-Kit@v6.0.0
  with:
    profile: github-level-1
    fail-on: fail

Action reference: docs/github-action.md. Starter workflows live under templates/workflows/.

Reports and Contracts

By default, evaluate writes reports/1.0 JSON. Older compatibility contracts and opt-in reports/2.0 are documented here:

• docs/reports-contract-v1.0.md
• docs/reports-contract-v2.0.md
• docs/sample-reports/

Exit codes:

Code	Meaning
0	Success; configured fail threshold was not violated
1	Evaluation completed and the fail threshold was violated
2	Usage, validation, or load error
3	Unexpected internal error

Supply Chain Verification

PyPI publication uses Trusted Publishing and registry attestations. Release artifacts also use GitHub Artifact Attestations. Container images are built from the checked-out release source tree, signed with cosign keyless, and attested.

Verification commands and limits are in docs/supply-chain-verification.md.

Documentation Map

Topic	Link
Documentation index	docs/README.md
Architecture	docs/architecture.md
CLI reference	docs/cli-reference.md
Results guide	docs/results-guide.md
Framework alignment	docs/framework-alignment.md
Positioning and limits	docs/positioning.md
EU CRA readiness	docs/cra-readiness.md
EU AI Act readiness	docs/eu-ai-act-readiness.md
MCP server security	docs/mcp-server-security.md
Release readiness	docs/release-readiness.md
Changelog	CHANGELOG.md

Repository Layout

Path	Purpose
src/oss_policy_kit/	Python package, CLI, evaluators, parsers, reporting
src/oss_policy_kit/data/	Bundled controls, profiles, and schemas
templates/	Starter workflows, waivers, docs, and ruleset examples
examples/	Hardened and vulnerable example repositories
tests/	Unit, application, integration, infrastructure, and property tests
docs/	User docs, architecture, mappings, ADRs, and release notes

Contributing and Security

• Contribution guide: CONTRIBUTING.md
• Governance: GOVERNANCE.md
• Vulnerability reporting: SECURITY.md
• Discussions: https://github.com/lucashgrifoni/OSS-Security-Policy-as-Code-Starter-Kit/discussions
• Issues: https://github.com/lucashgrifoni/OSS-Security-Policy-as-Code-Starter-Kit/issues

License

Apache-2.0. See LICENSE and NOTICE.