protein-structure-analysis

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill's primary function is to query authoritative scientific databases (AlphaFold and PDB) for protein structures. These are well-known, trusted academic resources maintained by organizations like EMBL-EBI and the wwPDB consortium.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection as it ingests untrusted data from external APIs (AlphaFold/PDB) and processes it into the agent context.
  • Ingestion points: Data is retrieved from external structural biology databases (SKILL.md).
  • Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potential instructions embedded in the scientific data.
  • Capability inventory: The skill is limited to data retrieval, structural alignment, and mapping variants; it does not perform dangerous operations like arbitrary code execution or filesystem modification.
  • Sanitization: No explicit sanitization of the retrieved database content is mentioned. This represents a standard, low-risk surface common to data-retrieval skills.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 01:03 PM
Security Audit — agent-trust-hub — protein-structure-analysis