skill-harness 0.24.3 → 0.25.1

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (3) hide show
  1. package/README.md +6 -0
  2. package/dist/index.js +11394 -2779
  3. package/package.json +3 -3
package/README.md CHANGED
@@ -59,3 +59,9 @@ each other.
59
59
  - Repo + full docs: https://github.com/mojomanyana/skill-harness
60
60
  - Step-by-step usage: [`docs/USAGE.md`](https://github.com/mojomanyana/skill-harness/blob/main/docs/USAGE.md)
61
61
  - CI Action (free spec lint on every PR): `uses: mojomanyana/skill-harness@latest` (or pin a version tag to freeze)
62
+
63
+ ## Optional decision evaluation and learning data
64
+
65
+ `skill-harness decision --help` exposes the npm-installed offline fixture, comparison, selected-session import and reviewed LoRA export workflow. `/skill-harness jev enable` asks for a fresh per-session storage choice every time; declining still permits explicitly confirmed JEV calls. Storage retains selected unlabeled inputs only and never authorizes training.
66
+
67
+ The exact Pi1.0.4 OAuth subscription comparison is separate from optional metered JEV. Reports preserve grouped splits, independent labels, abstention/error coverage and model/prompt identities. Synthetic fixtures qualify plumbing only. OpenAI Decisions remains excluded. See [the npm acceptance walkthrough](https://github.com/mojomanyana/skill-harness/blob/main/docs/DECISION-LEARNING.md) and the [data contracts](https://github.com/mojomanyana/skill-harness/blob/main/experiments/decision-shadow/LEARNING-WORKFLOW.md).