clawbio-guide

Pass

Audited by Gen Agent Trust Hub on Apr 5, 2026

Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill defines workflows that ingest untrusted data from local directories such as data/rnaseq/ and samples/ as described in SKILL.md. This configuration introduces a surface for indirect prompt injection where malicious instructions embedded in biological data files (e.g., in metadata or headers) could be processed by the agent.
  • [PROMPT_INJECTION]: Mandatory Evidence Chain for Indirect Prompt Injection: 1. Ingestion points: data/rnaseq/ and samples/ directories; 2. Boundary markers: None present in the documentation; 3. Capability inventory: Extensive subprocess calls for tools including BWA-MEM2, GATK, STAR, MACS2, and featureCounts; 4. Sanitization: No data validation or escaping mechanisms are defined.
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to install the @clawbio/clawbio plugin and references the project's official GitHub repository. These references are consistent with the skill's purpose and target the vendor's own namespace.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 5, 2026, 11:15 PM
Security Audit — agent-trust-hub — clawbio-guide