literature-review

Pass

Audited by Gen Agent Trust Hub on Apr 23, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of markdown instructions and references to external prompt templates. It does not include any executable scripts, shell commands, or configuration files that perform automated actions.\n- [INDIRECT_PROMPT_INJECTION]: The skill identifies an attack surface by ingesting external academic papers, though the risk is minimal in the context of research synthesis.\n
  • Ingestion points: External research papers and metadata retrieved by the researcher agent.\n
  • Boundary markers: None explicitly defined in this file.\n
  • Capability inventory: Writing synthesis reports and provenance logs to the outputs/ directory.\n
  • Sanitization: No specific sanitization or filtering logic is documented in this file.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 23, 2026, 09:11 AM