skill-evaluation-graphlisted

Deeply analyze, audit, score, and optimize agent skills conforming to the Agent Skills standard or Antigravity/Codex/Claude formats. Use when reviewing an existing skill, diagnosing why an agent misfires or burns context, pruning skill bloat, refactoring monolithic instructions into progressive disclosure, or benchmarking skill quality. Do not use for creating skills from scratch without an existing procedure (use workflow-skill-creator).
MaxLaurieHutchinson/skill-evaluation-graph · ★ 2 · AI & Automation · score 75

Install: claude install-skill MaxLaurieHutchinson/skill-evaluation-graph

# SEG: Skill Evaluation Graph: Procedural Driver for Agent Skill Audit & Optimization Turn agent skills into reliable, token-efficient, production-grade engineering assets. Evaluates trigger precision, progressive disclosure architecture, behavioral steering, execution determinism, operational safety, and token economics. --- ## Quick Reference Matrix | Concern | Purpose | Authoritative Resource | |:---|:---|:---| | **Canonical Terminology** | Authoritative vocabulary & domain definitions | [references/terminology.md](references/terminology.md) | | **6-Pillar Audit Rubric** | Objective 1–5 scoring definitions | [references/audit-rubric.md](references/audit-rubric.md) | | **Anti-Pattern Catalog** | Diagnostic guide for 14 recurring skill defects | [references/anti-patterns.md](references/anti-patterns.md) | | **Context Engineering** | Multi-tier progressive disclosure patterns | [references/progressive-disclosure-patterns.md](references/progressive-disclosure-patterns.md) | | **Evaluation Graph & Loop** | Autonomous Evaluator Loop Engine runbook | [references/workflow-graph-and-evaluator-loop.md](references/workflow-graph-and-evaluator-loop.md) | | **Harness Compatibility** | Cross-platform tool mappings (Antigravity/Claude/Codex) | [references/harness-tool-matrix.md](references/harness-tool-matrix.md) | | **Behavioral Trial Runner** | Control vs. Treatment evaluation & live trials | [scripts/eval_skill.py](scripts/eval_skill.py) | | **Capability Claim Audit** | Technical