dual-axis-skill-reviewer

You can automate skill code review by applying deterministic auto-axis checks and qualitative LLM review to produce reproducible quality scores. This dual-axis approach generates actionable improvement items and exports reports for merge gating or cross-project audits.

Dual-axis quality scoring combines deterministic code-based checks for structural validation with qualitative LLM review against a strict JSON schema. It merges weighted axes to produce reproducible, auditable quality scores for skill units.

Yes, you can emit LLM review prompts and accept strict JSON reviews against a schema to enforce quality thresholds. Configurable weight merging allows setting specific score gates, such as 90+, for CI integration and cross-project skill gating.

Configuring weight merging adjusts the contribution ratio between deterministic auto-axis checks and LLM deep review findings. This produces a final merged score that can be exported as JSON and markdown reports for tracking remediation.

The skill reviewer operates with zero dependencies, running optional test execution and artifact detection natively. It generates auto-axis metrics by validating script presence and execution safety directly within your repository environment.