bioinformatics
Warn
Audited by Gen Agent Trust Hub on Jun 16, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill provides instructions for the agent to clone external GitHub repositories (GPTomics/bioSkills and ClawBio/ClawBio) that are not part of the recognized trusted vendor list.
- [COMMAND_EXECUTION]: The skill recommends the installation of various bioinformatics tools and libraries using package managers such as pip, conda, brew, and apt. This includes the use of elevated privileges via sudo for system-level installations.
- [PROMPT_INJECTION]: The skill instructs the agent to read and follow instructions from external files (e.g., SKILL.md files in /tmp/bioSkills). This creates a surface for indirect prompt injection because external content is ingested into the agent context without sanitization or boundary markers.
- Ingestion points: Files located at /tmp/bioSkills///SKILL.md and /tmp/ClawBio/skills//README.md.
- Boundary markers: No specific delimiters or instructions to ignore embedded commands are present in the fetching workflow.
- Capability inventory: The environment supports file system operations, network access via git, and package/system tool installation.
- Sanitization: Content from the external repositories is cat-ed directly into the context without validation or filtering.
Audit Metadata