skills/freedomintelligence/openclaw-medical-skills/tooluniverse-antibody-engineering/Gen Agent Trust Hub
tooluniverse-antibody-engineering
Pass
Audited by Gen Agent Trust Hub on Jul 30, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill interacts with established scientific resources including IMGT (imgt.org), SAbDab (stats.ox.ac.uk), and IEDB (iedb.org) to retrieve biological sequence data, structural information, and immunogenicity records.
- [COMMAND_EXECUTION]: The provided Python implementation scripts facilitate the analysis workflow by calling platform-specific tools and performing file operations to save generated reports.
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting external data and user-provided sequences for inclusion in diagnostic reports.
- Ingestion points: Biological sequences (VH/VL) provided by the user and scientific metadata/epitope data fetched from external APIs (TheraSAbDab, IMGT, IEDB).
- Boundary markers: The workflow uses Markdown headers to separate analysis sections in the output report, but lacks explicit delimiters or instructions to ignore potential injection content within sequences.
- Capability inventory: The skill possesses capabilities for network communication via specialized scientific tools and file-writing for multiple report types (.md, .fasta, .csv).
- Sanitization: No explicit sanitization or validation logic is observed for the strings processed through the engineering pipeline before they are written to the local file system.
Audit Metadata