feat: add support-calibrated wording projection#337
Conversation
Stahl-G
force-pushed
the
codex/support-calibrated-wording-warnings
branch
from
528a0b1 to
429936c
Compare
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 429936c99a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| strong_wording = bool(_STRONG_WORDING_RE.search(context)) | ||
| framed = bool(_FRAMING_RE.search(context)) | ||
| attributed = bool(_ATTRIBUTION_RE.search(context)) | ||
| if support.get("blocking"): |
There was a problem hiding this comment.
For valid Claim-Support Matrix rows where support_label is unsupported, contradicted, or insufficient_evidence but the atom is not high-materiality and the row does not use block_release, the policy projection leaves blocking false. This check then suppresses unsupported_claim_reaches_reader and the human-review action even though support_labels contains an unsupported record, so status/Quality Panel miss the reader-facing risk this projection is meant to surface.
Useful? React with 👍 / 👎.
| return [], "claim_ledger_missing" | ||
| try: | ||
| payload = json.loads(path.read_text(encoding="utf-8")) | ||
| claims = ClaimLedger._claim_items_from_json(payload) |
There was a problem hiding this comment.
When claim_ledger.json is a list of objects with invalid ClaimContract fields, such as a blank statement/source_id or invalid v2 epistemic/source metadata, _claim_items_from_json only checks the container shape and still returns those entries. The projection can then report checked/pass instead of invalid_claim_ledger, causing status and Quality Panel consumers to lose the malformed-ledger signal for this diagnostic surface.
Useful? React with 👍 / 👎.
1 participant
Summary
Adds v0.11.2 PR3 Support-Calibrated Wording Warnings as a read-only Product OS projection.
support_wordingprojection over reader Markdown, Claim Ledger metadata, source taxonomy, and valid Claim-Support Matrix policy signalsstatus --json, human status, and Quality Panel / Summary / HTMLBoundary
This does not judge claim truth, generate or accept Claim-Support Matrix rows, run gates, block delivery, approve release, create repair authority, or create a quality score.
Validation
python3 -m pytest -q tests/test_support_wording.py tests/test_quality_panel.py tests/test_status.py tests/test_status_commands.pypython3 -m pytest -q tests/test_support_wording.py tests/test_quality_panel.py tests/test_status.py tests/test_product_baseline.py tests/test_release_consistency.py tests/test_status_commands.pypython3 -m pytest -qpython3 -m compileall -q src testspython3 scripts/check_skill_contract.pypython3 scripts/check_briefloop_skill_freshness.pypython3 scripts/sync_hermes_plugin_skills.py --checkpython3 scripts/check_release_consistency.py --no-tagpython3 scripts/check_version_consistency.pypython3 scripts/check_capabilities.pypython3 scripts/check_runtime_asset_parity.pypython3 scripts/generate_agent_configs.py --checkgit diff --check