Repository navigation
feat(benchmarks): enforce shared evaluation integrity policy - #57
Merged
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
New standard Harbor runs require a shared
restricted-v1policy for Nano, Pi, and DeepSeek Harness. Exact-host network allowlists are enforced outside the agent and checked before model work. All three adapters use process supervision, and collected output and network evidence remain on the host. Resource overrides are rejected, and non-root shared verifiers retain writable log directories.Completed outputs are frozen before evidence-bound review. Original rewards and errors remain intact; confirmed violations count as zero without leaving the denominator. Pending reviews prevent publication of adjusted pass rates. Historical reproductions require an explicit
--legacy-open-networkopt-out.The initial scope is single-step Linux Dockerfile/image tasks. Task images, official tests, and native budgets are preserved. Shared verifiers remain shared; the protocol does not establish complete protection against malicious submissions or reference exposure through allowed services.
Validation:
The English and Simplified Chinese Harbor READMEs are synchronized, and the 0.10.x changelog records the protocol under Unreleased. No development notes, research notes, or benchmark reports changed.