Re-verify stored claims on rerun and ship check-language - #17
Merged
Conversation
…d ship check-language in the wheel. Co-authored-by: Cursor <cursoragent@cursor.com>
|
| Filename | Overview |
|---|---|
| src/coldscreen/language.py | Adds shared evidence reconstruction and label-aware claim verification consistent with current deck and site evidence writers. |
| src/coldscreen/pipeline.py | Makes on-disk reruns pass sibling evidence into the memo language backstop and fail closed before writing. |
| src/coldscreen/check_language.py | Relocates the existing scanner into the installable package while reusing the shared verification helpers. |
| src/coldscreen/cli.py | Exposes the packaged scanner through a variadic check-language command and propagates its exit status. |
| scripts/check_language.py | Replaces the checkout-only implementation with a compatibility wrapper around the packaged scanner. |
| .github/workflows/ci.yml | Verifies that the installed wheel can run the language scanner against the packaged demo output. |
| tests/test_exemption_attacks.py | Adds regression coverage for tampered claims, missing evidence, wrong source labels, and valid verified quotations. |
Flowchart
%%{init: {'theme': 'neutral'}}%%
flowchart TD
A[Load stored case file] --> B[Load sibling evidence sections]
B --> C{Stored claim verifies against declared section?}
C -- Yes --> D[Allow claim quote as span exemption]
C -- No --> E[Do not exempt claim quote]
D --> F[Render and scan memo]
E --> F
F --> G{Language backstop clean?}
G -- Yes --> H[Write memo]
G -- No --> I[Fail without rewriting files]
Reviews (1): Last reviewed commit: "Honor a stored claim on rerun only after..." | Re-trigger Greptile
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
--render-onlyhonor a stored claim only after the sameclaim_quote_is_verifiedcheck CI uses, against siblingevidence_sections(AUD-003).--render-onlydoes not rewritecasefile.json.coldscreen check-language. The checkout script is a thin wrapper. The package job runs the installed command on the demo memo (AUD-006).Test plan
ruff check,ruff format --check,mypy,python scripts/check_language.pya fraudnot in evidence fails--render-only, and the golden puffery quote still passes