English static mirror for SEO/GEO · AI-assisted translation · Read Chinese original

SkillCoach: Process Auditing for Agent Skill-Use — A Paradigm Shift from Outcome Correctness to Process Quality

Forum topic · 小凯 · 2026-07-06

Summary

SkillCoach (arXiv:2607.01874) introduces a paradigm shift in AI agent evaluation: judging skill-use quality by process, not just final outcomes. The framework proposes a four-dimensional trajectory-level assessment covering skill selection, skill following, skill composition, and skill-grounded reflection, scored independently from external verifiers. A key contribution is detecting 'lucky successes' — trajectories where the agent reaches correct results through unreliable skill usage, such as selecting distractor skills, skipping critical steps, miscombining workflows, or omitting validation. SkillCoach uses a self-evolving rubric mechanism: gold-standard skills seed initial rubrics, evidence-based scoring and local arbitration iteratively refine weights and thresholds, and a validation gate accepts only updates that improve evaluation accuracy. Beyond assessment, the process scores serve as a training-data filter, prioritizing 'good process + correct result' trajectories and removing 'lucky success' samples that would otherwise reinforce faulty patterns. Experiments show this process-supervised filtering outperforms outcome-only data selection. The paper highlights applications in enterprise skill libraries containing semantically similar distractor skills, and discusses limitations including rubric dependency on skill definition quality, computational overhead, generalization to new skills, and integration with LLM-as-Judge methods.

Overview

Paper: SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use Authors: Jiayin Zhu, Kelong Mao, Yudong Guo, Dengbo He, Sulong Xu, Simiu Gu, Yutao Yue Link: https://arxiv.org/abs/2607.01874

TL;DR: Stop judging agents solely on task completion. SkillCoach audits the *process* of skill use across four dimensions, exposes 'lucky successes' hidden in current benchmarks, and doubles as a high-quality training-data filter.

---

The Blind Spot in Current Agent Evaluation

Most agent benchmarks rely on binary outcome metrics: task succeeded or failed. This works for simple settings but collapses in real skill-library environments because agents can reach correct results through four types of unreliable behavior:

1. Distractor skill selection — choosing a semantically similar but functionally wrong skill, then working around it 2. Skipped critical steps — bypassing required checks that happen to be unnecessary for the given input 3. Mis-composed workflows — wrong skill dependencies that don't break the specific case 4. Missing validation — skipping output verification that the external environment happens to forgive

All four patterns register as "success" under outcome-only evaluation, yet they encode unreliable behavior. Training on such trajectories bakes faulty skill-use patterns into the model.

---

Core Idea: Process Quality ≠ Outcome Success

SkillCoach reframes skill-use evaluation from outcome-oriented to process-oriented. The framework defines four trajectory-level meta-capabilities:

1. Skill Selection

  • Penalizes missed gold-standard skills
  • Penalizes selected distractor skills (semantically similar but functionally different)
  • Uses an F1-style score balancing precision and recall
  • 2. Skill Following

  • Verifies coverage of required execution steps
  • Penalizes skipped mandatory operations
  • 3. Skill Composition

  • Validates cross-skill dependency handling
  • Penalizes incorrect workflow chaining
  • 4. Skill-Grounded Reflection

  • Checks whether the agent performed skill-mandated output validation
  • Penalizes 'execute-and-exit' behavior with no verification
  • Key Design: Keep the External Verifier as an Independent Signal

    SkillCoach runs in parallel with, not as a replacement for, the external verifier:

  • Process score (SkillCoach's four dimensions)
  • Outcome signal (external verifier: success/failure)
  • This yields a 2×2 quadrant:

  • ✅ Good process + correct outcome → high-quality sample
  • ⚠️ Poor process + correct outcome → 'lucky success' (dangerous)
  • ⚠️ Good process + wrong outcome → promising sample (possible environmental noise)
  • ❌ Poor process + wrong outcome → clear error sample
  • Traditional evaluation conflates the first two rows. SkillCoach isolates the second — the most harmful one for training.

    ---

    Self-Evolving Rubric Mechanism

    Step 1: Skill-Dependent Task Filtering

    Not every task meaningfully tests skill use. A three-layer filter excludes non-skill-dependent tasks:

    1. No-skill success rate — if the agent succeeds without any skill, the task doesn't depend on skills 2. Gold-skill gain rate — does using the correct skill significantly raise success rate? 3. Critical-step coverage — does the task actually exercise the skill's key steps?

    Step 2: Initial Rubric Generation from Gold Skills

    From filtered gold-standard skills, the framework auto-extracts:

  • Skill names and descriptions
  • Critical step lists
  • Dependency graphs
  • Output validation requirements
  • These seed the initial four-dimensional scoring rules.

    Step 3: Validation-Gated Self-Evolution

    Initial rubrics are imperfect. Three sub-steps drive iterative refinement:

    1. Evidence-based scoring — run rules on real interaction trajectories, record per-dimension scores and deductions 2. Local arbitration correction — compare process scores against outcome signals; investigate mismatches to fix weights and thresholds 3. Validation gate — new rubrics must pass consistency checks on a held-out set; if they produce more misjudgments, revert to the prior version. Only accuracy-improving updates are retained

    This mechanism is what enables adaptation to messy, real-world enterprise skill libraries full of overlapping and distractor skills.

    ---

    Dual Value: Evaluation + Training Enhancement

    SkillCoach doubles as a training-data quality filter.

    Process Supervision > Outcome Supervision

    Traditional filtering:

  • Keeps outcome-successful trajectories (contaminated with 'lucky successes')
  • Discards outcome-failed trajectories (may include strong processes masked by environment noise)
  • SkillCoach's process-supervised filtering:

  • Prioritizes 'good process + correct outcome'
  • Removes 'lucky successes' that would amplify faulty patterns
  • Retains 'good process + wrong outcome' samples for failure analysis
  • Provides rich multi-dimensional supervision signals
  • Experiments report that data filtered by SkillCoach yields significantly stronger skill-use capabilities than outcome-only filtering.

    ---

    Why It Matters

    1. From black-box scoring to white-box auditing — developers see exactly where in the skill-use pipeline the agent underperforms 2. Exposes the 'lucky success' risk — the first systematic method to surface this hidden failure mode 3. Enterprise-skill-library ready — designed for noisy environments with overlapping and distractor skills 4. Evaluation-as-training — process scores are directly usable as training supervision, enabling a closed 'evaluate → find errors → improve training → re-evaluate' loop

    ---

    Limitations and Open Questions

    1. Rubric quality depends on skill definitions — unclear skill descriptions degrade the initial rubric; self-evolution partially compensates but requires sufficient trajectories 2. Computational overhead — step-level four-dimensional analysis is more expensive than simple outcome checks 3. Generalization to new skills — adding a novel skill requires new trajectories and rubric updates 4. LLM-as-Judge integration — current rubrics are structured; combining with LLM-as-Judge flexibility for more open-ended, semantic skill use is an open direction

    ---

    One-Sentence Summary

    Agent skill-use quality cannot be defined by task outcomes — it must be defined by process correctness. Through four-dimensional process auditing (selection, following, composition, reflection) and validation-gated self-evolution, SkillCoach detects hidden 'lucky successes' and produces higher-quality training data than outcome-only filtering, marking a key step from black-box scoring toward white-box auditing of AI agents.

    ---

    Reference

  • Paper: SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use (arXiv:2607.01874)
  • Authors: Jiayin Zhu, Kelong Mao, Yudong Guo, Dengbo He, Sulong Xu, Simiu Gu, Yutao Yue
  • Core framework: Self-Evolving Rubrics
  • Four dimensions: Skill Selection, Skill Following, Skill Composition, Skill-Grounded Reflection
  • Technical pillars: Skill-dependent task filtering protocol + validation-gated self-evolution
  • Dual value: Evaluation quality uplift + training-data filtering
  • Key design: Decoupling process score from outcome signal
  • Experimental finding: Evolved rubrics simultaneously improve evaluation quality and downstream training
  • Target scenario: Enterprise skill libraries with distractor skills

Tags

#ai-agents#process-evaluation#self-evolving-rubrics#skill-use#training-data-filtering#llm-evaluation#arxiv-2607-01874#process-supervision

This page is an English static mirror generated for search and AI citation. It may be a full translation or structured summary of the Chinese original. Canonical interactive discussion lives on the Chinese page: https://zhichai.net/topic/178209089