Skip to content

test(drive): validate AI evidence extraction through the executable - #86

Merged
leemour merged 1 commit into
mainfrom
test/drive-ai-extraction
Oct 15, 2026
Merged

leemour merged 1 commit into
mainfrom
test/drive-ai-extraction

Conversation

@leemour

@leemour leemour commented Oct 15, 2026

Copy link
Copy Markdown
Contributor

Drive evidence extraction requires exact quotes, while the shared model gateway forbids quoting source data. Add a reproducible executable-level regression and guarded loopback model responder. Record npm 0.12.0 as the known-conflict baseline and the local Core/Drive source-quote candidate separately, including candidate module hashes; no released fix is claimed.

Both runs pass 59 CLI commands and 12 loopback provider requests across preview, scoped proposals, confirmation, tab isolation, malformed/error responses, network guards and a real sync while extraction is in flight. This does not measure real model quality. No owner store, Google account, model credential or real document is used.

Validation: both recorded runs, targeted Biome and docs:check pass. Shared API fix: WireCatLabs/cli-core#1317. Published-package validation will be recorded after the coordinated Core and Drive releases.

@leemour
leemour merged commit ce26369 into main Oct 15, 2026
1 check passed
@leemour
leemour deleted the test/drive-ai-extraction branch October 15, 2026 21:20
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