Adityaraj0421/naksha-studio v4.2.0

v4.2.0 — Eval Quality & CI Hardening

on GitHub

v4.2.0 — Eval Quality & CI Hardening

Eval quality

• Upgraded v4.1.0 tutorial evals (IDs 140–145) from 3-4 shallow keyword checks to rich 6-assertion format
• All 146 evals now follow consistent named-assertion pattern

CI hardening — 3 new checks (5 → 8 total)

• Check 6: pipeline-yaml-structure — validates all pipeline YAML files have name, description, steps, command fields
• Check 7: skill-command-sync — ensures every command in commands/ is referenced in skills/design/SKILL.md
• Check 8: command-allowed-tools — enforces allowed-tools: frontmatter on all 57 command files; fixed 17 commands that were quietly missing it

Stats

• Roles: 26 · Commands: 57 · Knowledge lines: 13,500+ · Reference files: 32
• Evals: 146 · CI checks: 8 · Pipelines: 7