菜单
Load when an agent, skill, or prompt needs strict review. Focuses on failure modes, severity, and clear feedback.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Load when writing or reviewing agent .md files. Covers agent artifact shape, frontmatter, and Meridian-specific agent conventions.
Load when writing or reviewing prompts, skills, or agents. Gives the core heuristics for writing prompts that steer LLM behavior reliably.
Load when writing or reviewing skills. Covers SKILL.md shape, loading mechanics, and how to structure bundled resources.
| name | prompt-review |
| type | reference |
| description | Load when an agent, skill, or prompt needs strict review. Focuses on failure modes, severity, and clear feedback. |
| model-invocable | false |
Review prompt quality first. Use /prompt-principles and /llm-writing as the default lens.
Read all four /prompt-principles resource files before reviewing. The body routes; the resources carry the actual heuristics:
resources/prompt-level.md: attention, structure, leading words, heuristic altituderesources/skill-level.md: loading model, extraction, decompositionresources/agent-level.md: cognitive mode, body focus, reuseresources/system-level.md: handoffs, verification, coordination altitudeReview against each layer that applies. Focus on high-leverage issues: no-op lines, misplaced knowledge, overloaded bodies, weak routing, brittle prescription, and missing reasons. A good finding points to the exact text, explains the failure mode, and suggests a better direction.
After the prompt-quality pass, do a lighter mechanics pass when relevant. For that second angle, load resources/mechanics.md.
Start with a short overall assessment. Then list the findings and include the severity of each. End with a verdict: approve, approve with notes, or request changes.