Skip to content

feat: enforce regulated content boundaries - #3

Merged
salian merged 1 commit into
mainfrom
codex/release-hardening
Aug 8, 2026
Merged

feat: enforce regulated content boundaries#3
salian merged 1 commit into
mainfrom
codex/release-hardening

Conversation

@salian

@salian salian commented Aug 8, 2026

Copy link
Copy Markdown
Owner

Summary

  • infer regulated and YMYL signals from public bundle metadata
  • backfill explicit professional-review boundaries across the public catalog
  • make enrichment add missing risk metadata and validation reject omissions
  • exercise the policy in repository CI

Verification

  • canonical validator: 0 errors, 18 pre-existing relationship warnings
  • content-risk unit tests: 4 passed
  • all 277 public compliance records have explicit risk metadata
  • zero public inferred-risk records are missing content_risk
  • semantic registry diff: 323 risk additions and four word-count refreshes

SEO impact

No sitemap, robots, canonical, or indexability-policy changes.

What changed:
- Added conservative regulated and YMYL inference from bundle category, title, aliases, and tags.
- Backfilled explicit professional-review boundaries on 323 public registry records.
- Made enrichment add inferred risk metadata and canonical validation reject detected risk signals without it.
- Added inference regression tests to the public repository validation workflow.
- Ignored generated Python bytecode artifacts.

Why:
- Consequential legal, medical, financial, employment, privacy, security, safety, and compliance content must not rely on editors remembering a manual warning.
- Explicit metadata gives the website a stable, reviewable professional-reliance boundary for every detected public record.

Verification:
- Automatic inference tests passed.
- Canonical validation completed with 0 errors and 18 pre-existing relationship warnings.
- All 277 public compliance bundles have explicit risk metadata.
- No public record with an inferred risk signal is missing content_risk metadata.
- Semantic registry review found only 323 risk additions and four truthful word-count refreshes.

Notes/Risks:
- Inference is intentionally conservative and may display a not-reviewed boundary on some tool or operational pages whose use can affect consequential decisions.
- The 35 excluded draft records are not backfilled until they become publication candidates.
@salian
salian merged commit a350ddd into main Aug 8, 2026
1 check passed
@salian
salian deleted the codex/release-hardening branch August 8, 2026 09:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant