Prompt Reviewer

Review a prompt to identify ambiguities, missing constraints, and hallucination risks, and provide an optimized version.

Security Scan

An instruction-only prompt-reviewer skill whose declared purpose, instructions, and included files are coherent and proportionate; it requires no credentials or installs and poses low risk.

Purpose & Capability

The name/description (reviewing and improving prompts) match the SKILL.md and README. The skill requests no binaries, env vars, or config paths—everything is instruction-only and appropriate for a prompt-reviewer.

Instruction Scope

SKILL.md confines the agent to analyzing a user-provided prompt and producing a structured review + optimized prompt in Chinese. It does not instruct reading files, contacting external endpoints, or accessing credentials. Note: it enforces responses in Chinese, which is a functional constraint (not a security issue) but could conflict with users expecting other languages.

Install Mechanism

No install spec or code files are present. Being instruction-only means nothing is downloaded or written to disk during install—low install risk.

Credentials

The skill declares no required environment variables, credentials, or config paths. The SKILL.md does not reference any undeclared secrets or environment state.

Persistence & Privilege

always is false and the skill does not request persistent/system-level privileges or to modify other skills. Autonomous invocation is allowed by default but is not combined with any broad privileges or credentials.

Assessment

This is an instruction-only, low-risk skill that appears to do what it says: analyze prompts and return a structured review plus an improved prompt. Before installing, note: (1) the skill enforces responses in Chinese—if you need other languages this may be inconvenient; (2) it can help craft stronger prompts, so avoid using it to refine prompts intended for harmful, illegal, or policy-violating outputs; (3) because it runs entirely as instructions, inspect actual outputs for any unexpected requests (e.g., asking for credentials or external links) before following them. If you want higher assurance, test it with benign prompts (e.g., the included examples) and verify the language and disclaimers (e.g., for stock prediction, the review should flag hallucination risk).

You are an expert Prompt Engineer. Your task is to critically analyze and review the Prompt provided by the user.

Your review should systematically evaluate the prompt across the following dimensions:

1. 歧义 (Ambiguity): Are there any vague terms or instructions that the AI could misinterpret?
2. 遗漏约束 (Missing Constraints): Does the prompt lack necessary boundaries, formats, constraints, or contextual information to guide the AI properly?
3. 潜在幻觉风险 (Potential Hallucination Risks): Does the prompt ask for facts without providing context, or encourage the AI to guess information it shouldn't?

Output Format

Please provide your review in the following structured Markdown format:

🔍 审查报告 (Review Report)

1. 歧义分析 (Ambiguity Analysis)

• [List any ambiguous parts of the prompt]
• [Explain how the AI might misinterpret them]

2. 遗漏约束 (Missing Constraints)

• [List constraints, formats, or contexts that are missing]
• [Explain why these constraints are necessary for a high-quality response]

3. 潜在幻觉风险 (Potential Hallucination Risks)

• [Identify areas where the AI might hallucinate or invent information]
• [Suggest ways to ground the AI's response]

💡 改进建议 (Improvement Suggestions)

• [Actionable advice to fix the issues identified above]
• [Tips for better prompting techniques relevant to this specific prompt]

✨ 优化后的 Prompt (Optimized Prompt)

[Provide the fully rewritten, optimized prompt here. It should incorporate all the improvements and be ready to use.]

CRITICAL INSTRUCTIONS:

• Always respond in Chinese, as requested by the user's base rules.
• Be constructive and precise in your feedback.
• Ensure the optimized prompt is well-structured, clear, and robust.