OIDF publication-checks

Validation tools for the OpenID Foundation specification publication pipeline.

Performs automated checks on proposed specifications submitted via pull requests to the openid/publication repository.

Structure

• spec_validator.py - Core validation library. Regex-based HTML analysis for spec structure, state detection, dates, notices, references, authors, history. Also works as CLI (python3 spec_validator.py -content-state file.html).
• process.py - Pre-publication validation. Runs all checks against proposed specs. Called by GitHub Actions on propose/** branches.
• publish.py - Publication file preparation. Copies spec files to to-publish/ with correct naming for DRAFT/FINAL/ERRATA/DRAFT_ERRATA states. Called by GitHub Actions on merge to main.

Running Tests

pip install -r requirements.txt
python3 -m pytest tests/ -v                    # all tests
python3 -m pytest tests/ -m "not e2e"          # unit tests only
python3 -m pytest tests/ -m e2e                # e2e tests

E2e tests create temporary git repos and run process.py/publish.py against them. Network access is not required — CSV fetching is skipped via SKIP_CSV_FETCH and URL accessibility checks are mocked via MOCK_URL_CHECK (both set automatically by the test helper).

Checks Performed

See CLAUDE.md for a detailed list of all checks and the OIDF rules they enforce.

Reference URL checks fall back to the Internet Archive Wayback Machine when a site blocks automated requests (HTTP 403/429, or a 202 AWS WAF challenge): a snapshot means the URL passes, no snapshot means it fails, and an unreachable archive produces a non-blocking warning.

See ERROR-MODES.md in the publication repo for guidance on fixing each failure.