Repository navigation
chore(release): prepare v0.5.0 - #185
Merged
Merged
Conversation
Signed-off-by: Narendran Raghavan <nraghavan@nvidia.com>
rng1995
requested review from
chrisknvidia,
mosheabr and
rmalani-nv
as code owners
October 6, 2026 17:25
rng1995
enabled auto-merge
October 6, 2026 17:29
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Prepare SkillEvaluator v0.5.0 after merging the Harbor 0.24.0 upgrade in #83, following the v0.4.0 release PR (#177).
Bump package metadata, the source version fallback, citation metadata, the developer guide example, and the root lockfile entry to
0.5.0. Cut the dated0.5.0 - 2026-10-06changelog section and leave a fresh Unreleased section. Dependency versions and historical release notes are unchanged.Release Impact
This release includes the changes merged since v0.4.0: Harbor 0.24.0 (#83), configurable judge evidence budgets (#160), structured judge output and transient-error retries (#145), trajectory reconstruction from agent logs (#119), canonical native-task case identity (#161), and non-string YAML-key validation (#178).
The release notes cover Harbor's upgrade requirements, including numeric native-task rewards,
wandbusingcwsandbox, disabled project dotenv loading, and separate-verifier precedence. This PR prepares the release metadata; thev0.5.0tag and GitHub Release can be published against its merge commit after validation.Verification
CHANGELOG.md; no credentials, private datasets, or proprietary benchmark content added.uv lock --checkanduv sync --all-extras --locked --python 3.13.make lintandruff check ..check: 0 errors; authenticated redirect check skipped withoutFERN_TOKEN.make build: wheel and source distribution built successfully.twine==6.2.0 check --strictpassed for both distributions.make testrun on Python 3.13 with four workers: 11,250 passed, 20 skipped, and one timing-sensitive failure intest_bounded_harbor_process_timeout_preserves_redacted_diagnostic_tail(0.1-second subprocess deadline; expected diagnostic was absent). The entire affected file passed sequentially: 100 passed, 2 skipped. No test or runtime changes were made.0.5.0, with Harbor0.24.0; imports, bundled verifier, progress, and CLI smoke checks pass.No new behavioral tests are needed for this metadata-only change; existing tests validate the release.