Repository navigation
feat(harbor): add pi agent baselines and audit dev34 results - #54
Merged
Merged
Conversation
minixalpha
changed the title
minixalpha
marked this pull request as ready for review
October 30, 2026 15:06
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The standard Harbor workflow now runs a pinned Pi distribution with explicit model registration, preserved native artifacts, and auditable usage. Independent configuration snapshots and verified concurrent cache publication prevent job inputs from being replaced. Process supervision confirms cancellation before grading, and failed streams retain partial accounting.
Original task images can be cached with verified ranged downloads, registry-token renewal, native OCI metadata, and bounded batch storage. Original package sources, task resources, and official tests are preserved.
The existing 34-task development pool received one Pi + DeepSeek Flash attempt per task: 26 official passes, six failures, and two unscored legacy timeouts. The bilingual reports and audited JSON preserve raw exceptions, source amendments, public-reference exposure, credential-output findings, and partial costs. No complete, uncontaminated 34-task pass rate is claimed; raw traces remain private.
Validation: 269 Harbor tests pass after integration with current main. Frozen-output hashes and the synchronized English/Chinese report pair verified. The four Pi workflow review findings have regression coverage. The changelog records the new baseline workflow.