Repository navigation
v0.3.0 — single-file task config + per-dimension scoring
Changed
- Breaking: per-task config consolidated into a single
task.toml. The previous split betweentask.json(nasde-only) andtask.toml(Harbor) is gone — Harbor sections ([task],[agent],[environment],[verifier],[metadata]) and nasde-specific fields ([nasde.source]) now live side by side in one file. All bundled examples migrated. (#29) - Breaking:
modelis now required in everyvariant.toml. There is no project-level model default innasde.tomlanymore — different agent families need different models, so each variant declares its own. Missing model fails fast with a clear error. CLI--modelstill overrides. (#29)
Added
- Independent per-dimension scoring scales (ADR-008). Each dimension in
assessment_dimensions.jsondeclares its ownmax_score(any positive integer).normalized_scoreis computed from the actual sum of per-dimension maxima — pick the granularity that matches what you can distinguish, instead of forcing every benchmark into a fixed total. (#34)
Fixed
- Evaluator now prints a clear, actionable error when the underlying
claudeorcodexCLI is missing, instead of failing with a confusing subprocess trace. (#30) - The "missing CLI" hint now prints
[evaluation]correctly instead of hiding it as Rich markup.
Security
- Pin
litellm>=1.83.7to address GHSA-xqmj-j6mv-4862. (#31)
Docs
- README: clarified how NASDE is meant to be used end-to-end.
- README: Harbor links now point at
harborframework.cominstead of the GitHub repo, matching Harbor's primary site.
Internal
CODEOWNERSadded so PRs auto-request review from the right people.
Install: uv tool install git+https://github.com/NoesisVision/nasde-toolkit.git@v0.3.0
Full changelog: v0.2.1...v0.3.0