| name | evaluate-matrix |
| description | Cross-model skill evals with real execution and grading — the executability gate. Use when checking whether a weak model can actually do a skill, not just comprehend it. |
| args | <plugin/skill-name> [--models opus,haiku] [--with-skill-only] [--runs N] |
| allowed-tools | Task, Read, Write, Edit, Glob, Bash(bash *), Bash(python3 *), TodoWrite |
| argument-hint | git-plugin/git-commit --models opus,haiku --with-skill-only |
| model | opus |
| agent | general-purpose |
| created | "2026-06-13T00:00:00.000Z" |
| modified | "2026-07-18T00:00:00.000Z" |
| compatibility | claude-code |
| reviewed | "2026-06-13T00:00:00.000Z" |
/evaluate:matrix
The executability gate. Where /evaluate:legibility asks "can a fresh
agent comprehend this skill," this skill asks "can a weak model actually
do it" — run the skill's evals on haiku (and opus/sonnet) with real tool
execution, grade the produced artifact, and surface the per-skill verdict
executable_on_haiku. A skill that opus passes and haiku fails leans on
reasoning the cheap model lacks.
This builds the orchestration the cross-model design
(docs/cross-model-evaluation.md)
calls a follow-up. It reuses, without duplicating: prepare_run.sh,
grade_deterministic.py (zero-token first pass), the eval-grader agent
(deferred fuzzy checks only), model-matrix.json, and render_matrix_report.py.
When to Use This Skill
| Use this skill when... | Use alternative when... |
|---|
| Checking whether a weak model can execute a skill | Checking whether the SKILL.md reads clearly -> /evaluate:legibility |
| Running the Tier-2 cross-model sweep on a golden-set skill | Single-skill, single-model effectiveness -> /evaluate:skill |
| Diagnosing a skill that opus passes but haiku fails | Structural/lint validation -> scripts/plugin-compliance-check.sh |
| Re-checking canaries after a new model ships | Improving a skill from results -> /evaluate:improve |
Context
- Available plugins: !
find . -maxdepth 2 -type d -name '*-plugin' -not -name '.claude-plugin'
Parameters
| Parameter | Default | Description |
|---|
<plugin/skill-name> | required | Target skill as plugin-name/skill-name |
--models <list> | opus,haiku | Comma-separated pinned aliases to run |
--with-skill-only |
