SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use
For developers of LLM agents using skill repositories, this addresses the problem of coarse final-verifier evaluation that misses process errors, offering a more granular and self-improving evaluation and training signal.
SkillCoach introduces a self-evolving rubric framework that evaluates LLM agent skill-use along four process dimensions (selection, following, composition, reflection), improving evaluation quality and providing stronger training supervision than outcome-only methods.
Skills are becoming a reusable operational layer for LLM agents, encoding SOPs, domain rules, tool workflows, scripts, and validation routines. In realistic skill repositories, overlapping skills make reliable skill-use difficult. Final verifier success is too coarse for both evaluation and training, since an agent may pass through trial and error while selecting distractor skills, skipping required steps, composing workflows incorrectly or omitting final checks. We introduce SkillCoach, a self-evolving rubric framework for evaluating and enhancing agentic skill-use. SkillCoach derives skill-grounded process rubrics from real rollouts and evaluates trajectories along four dimensions: skill selection, skill following, skill composition, and skill-grounded reflection. It keeps the external verifier as a separate outcome signal, allowing process quality to be distinguished from accidental task success. The evolved rubrics further serve as process supervision for selecting high-quality training trajectories. Experiments show that evolved rubrics substantially improve evaluation quality, expose failures hidden by final accuracy, and provide stronger supervision signals than outcome-only filtering for enhancing agentic skill-use.