malware-analysis

Pass

Audited by Gen Agent Trust Hub on Mar 7, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: Indirect Prompt Injection Surface
  • Ingestion points: The skill (SKILL.md) ingests untrusted data through malware samples and YARA rules provided via the CLI and API endpoints.
  • Boundary markers: The documentation lacks mention of delimiters or instructions to ignore potential commands embedded within analyzed samples.
  • Capability inventory: The skill (SKILL.md) performs network operations to vendor and well-known analysis services, reads local files, and executes commands through the CLI.
  • Sanitization: There is no evidence of sanitization for the strings or behavioral data extracted from samples before they are processed by the agent.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 7, 2026, 05:44 PM