book-finder

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill explicitly instructs constructing and running curl commands that include a URL query parameter "&key=KEY", which requires the agent to capture and embed a fetched key value verbatim into command output (an insecure pattern of passing secrets on the command line).

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). SKILL.md requires runtime web searching and scraping of third-party sites (e.g., Libgen/Anna’s Archive/Archive.org/university servers) and parsing their returned HTML to extract download links, which is outsider-authored free text ingested into the agent’s context via the fetched page content.