fix: decode quoted benchmark shell commands - #459
Merged
Conversation
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Description
The benchmark can reject a valid run when an agent searches for
run:android|agent-device|emulator. Codex's displayed shell command escapes the inner quotes; stripping only the outer quotes makes the parser split that search pattern into fake commands.Solution
Decode one outer double-quoted shell argument before auditing, using the same normalizer for command segments and proof matching. Literal backslashes retain shell quoting semantics. Standalone proof boundaries and session requirements remain unchanged. This is collector-only: no CLI, prompt, or timing changes.
Test plan
Regression coverage uses the encoded quoted-search reproduction, quoted proof text, chained proof rejection, and literal backslashes. The standalone-proof self-test passes. Re-collection of the affected retained run clears the false positive with identical evidence hashes and timing; its original audit remains preserved. Existing real setup violations still fail the corrected audit.
Fixes #458.