dag-skills-matcher

Match task descriptions to suitable skills using semantic similarity and performance history. The system normalizes input, scores candidates by fit and Elo, and filters results with NOT clauses to streamline DAG node assignments.

Rank candidate skills using a multi-criteria scoring system that evaluates semantic similarity, Elo ratings, and historical performance. This ranking identifies the best-fitting skills and routes escalations when suitable matches are not found.

Semantic similarity compares normalized natural language task descriptions against a browsable skill catalog. It calculates fit scores to identify and retrieve the most compatible skills available for your automation workflow.

Yes, you can exclude specific skills using NOT clause filtering. This removes unwanted candidates from the matching results before final ranking, ensuring only relevant skills are considered for your task.

Use escalation routing when semantic similarity and performance history cannot identify a suitable skill match for your task description. It handles edge cases where no existing catalog entry meets the required fit score threshold.

No dependencies are required. The skill matcher operates independently to maintain a browsable skill catalog, providing input normalization, multi-criteria scoring, and candidate ranking for task matching without external libraries.