research-topic-extractor
Pass
Audited by Gen Agent Trust Hub on Jun 18, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary function is to analyze academic content from multiple sources, such as files, URLs, and images, to extract research topics, keywords, and core questions. This behavior is consistent with its stated purpose and does not involve any malicious activities.
- [SAFE]: The skill handles diverse input formats including PDF, Word, and Markdown using the agent's native file-reading tools and LLM vision capabilities. No unauthorized command execution, privilege escalation, or persistence mechanisms were detected.
- [SAFE]: Integration with external tools like
ghand thebensz-collect-bugsskill is clearly documented for the purpose of reporting bugs to the author's public repository (huangwb8/bensz-bugs). This is a legitimate vendor resource functionality triggered by explicit user requests. - [SAFE]: The prompt templates provided in the skill use clear structural delimiters (e.g.,
【输入内容】) to separate user-provided data from the extraction instructions. This design helps the agent distinguish between content to be summarized and instructions to be followed, which mitigates common indirect prompt injection risks.
Audit Metadata