# Prompt Review

> Load when an agent, skill, or prompt needs strict review. Focuses on failure modes, severity, and clear feedback.

- Skill: `haowjy/prompt-review` (Agent Skill, multi-file: 2 files)
- Install (CLI): `npx skillmds@latest add haowjy/prompt-review`
- Raw SKILL.md: https://api.skillmd.com/api/skills/haowjy/prompt-review/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: haowjy (https://skillmd.com/u/haowjy)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/haowjy/prompt-review

---


# Prompt Review

Review prompt quality first. Use `/prompt-principles` and `/llm-writing` as the default lens.

## What to Look For

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 altitude
- `resources/skill-level.md`: loading model, extraction, decomposition
- `resources/agent-level.md`: cognitive mode, body focus, reuse
- `resources/system-level.md`: handoffs, verification, coordination altitude

Review 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`.

## Report

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.