Skip to content

fix(review): preserve directly established findings - #153

Merged
morgaesis merged 1 commit into
mainfrom
codex/v0.8.21-benchmark-regression
Aug 16, 2026
Merged

morgaesis merged 1 commit into
mainfrom
codex/v0.8.21-benchmark-regression

Conversation

@morgaesis

Copy link
Copy Markdown
Contributor

Directly established defects must remain publishable when the model also speculates about callers or consumers. This makes the initial and repair prompts omit repositoryContext unless the conclusion truly depends on repository-wide search, and records why an authored target was suppressed without changing benchmark scoring.

@morgaesis
morgaesis enabled auto-merge (squash) August 16, 2026 03:40
@morgaesis
morgaesis merged commit 0467a7d into main Aug 16, 2026
10 checks passed
@morgaesis
morgaesis deleted the codex/v0.8.21-benchmark-regression branch August 16, 2026 03:45
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