alterlab-alphafold
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the Bash tool to execute colabfold_batch for structure prediction.
- [DATA_EXFILTRATION]: By default, the skill sends protein sequences to a hosted MMseqs2 server; this data flow is disclosed in the skill's reference documentation for MMseqs2-accelerated MSAs.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted protein sequences from external FASTA files, representing an attack surface for indirect prompt injection. * Ingestion points: input.fasta and complex.fasta referenced in SKILL.md. * Boundary markers: Absent. * Capability inventory: Bash, Read, Write, and Edit tools across SKILL.md. * Sanitization: Absent in instructions.
Audit Metadata