radio
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill describes functionality for reading messages from other nodes, which introduces an attack surface for indirect prompt injection from external sources.\n
- Ingestion points: Data is ingested via
fractal radio read,fractal radio messages, andfractal radio feedcommands described in SKILL.md.\n - Boundary markers: The skill includes explicit instructions for the agent to treat incoming feed content as information to weigh rather than instructions to adopt, emphasizing that its primary binding rules are its seed (NODE.md, steps, skills).\n
- Capability inventory: The agent uses the
fractal radioCLI to read, write, and relay communications.\n - Sanitization: The documentation provides specific warnings to the agent about shell command substitution risks (e.g., backticks and dollar signs) in message bodies and recommends using Python argv for safer message composition.
Audit Metadata