laorenfang-zhuangxiu-zhinan
Pass
Audited by Gen Agent Trust Hub on Nov 3, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill package is composed strictly of Markdown files (SKILL.md and 206 reference files). There are no Python scripts, Node.js modules, or binary executables included.
- [PROMPT_INJECTION]: Content analysis of SKILL.md and the reference files revealed no behavioral overrides, safety bypasses, or persona-based jailbreak attempts. The instructions are limited to providing renovation advice in a specific human-like tone.
- [DATA_EXFILTRATION]: No patterns related to harvesting credentials, accessing sensitive system paths (~/.ssh, ~/.aws), or unauthorized data transmission to remote servers were detected.
- [OBFUSCATION]: The skill's text was analyzed for Base64 encoding, zero-width characters, and homoglyphs. No obfuscation techniques used to hide malicious intent were found.
- [INDIRECT_PROMPT_INJECTION]: The skill provides a task-specific knowledge base for RAG (Retrieval-Augmented Generation). It does not request or require access to dangerous tools like shell execution or file system modification, making the data ingestion surface safe.
Audit Metadata