bioinformatics

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly instructs the agent to git-clone and read public GitHub repos (https://github.com/GPTomics/bioSkills and https://github.com/ClawBio/ClawBio) and to "read the specific skill" files (SKILL.md/README), meaning the agent will ingest untrusted, user-generated third-party content that can directly influence tool use and execution.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs runtime git clones of https://github.com/GPTomics/bioSkills and https://github.com/ClawBio/ClawBio and then directs installing requirements and running ClawBio pipelines (pip install -r requirements.txt, demo scripts), so fetching those repos at runtime can supply and lead to execution of remote code.