kegg-database
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [METADATA_POISONING]: The skill's YAML frontmatter lists 'K-Dense Inc.' as the author, which differs from the system-reported author identifier 'wu-yc'. While potentially a branding choice, this metadata inconsistency is noted.\n- [INDIRECT_PROMPT_INJECTION]: The skill facilitates the retrieval and processing of data from the external KEGG REST API. Processing such external data without defining strict boundary markers or sanitization logic creates a potential surface for indirect prompt injection.\n
- Ingestion points: Biological data entries (pathways, genes, compounds) fetched from the KEGG API through functions in 'scripts/kegg_api.py'.\n
- Boundary markers: The skill instructions do not specify the use of delimiters or 'ignore' instructions for the fetched content.\n
- Capability inventory: The skill is designed for network-based data retrieval and subsequent processing in Python workflows.\n
- Sanitization: No specific sanitization or validation mechanisms for external API responses are described in the provided instructions.
Audit Metadata