update-agent
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions involve taking user-provided input for agent configurations (model names, descriptions, and IDs) and interpolating them into shell commands. This creates a potential surface for prompt injection if the inputs are not properly sanitized before being passed to the tool.
- Ingestion points: SKILL.md (prompts for
agent_id,model, anddescription). - Boundary markers: Absent; inputs are placed directly into JSON payloads.
- Capability inventory: Network communication and proxy management via
curl(Bash). - Sanitization: Not explicitly defined within the skill instructions, relying on the agent's default handling of tool arguments.
Audit Metadata