Skills
skills/hillstone-networks/agent-skills/agent-roles-and-subagents/Gen Agent Trust Hub

agent-roles-and-subagents

Pass

Audited by Gen Agent Trust Hub on Mar 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a local Python script, scripts/validate_roles.py, to ensure generated JSON configurations match the expected schema. Evidence: Manual execution instructions in SKILL.md. Capability: The script reads local files and performs JSON parsing. This is a standard vendor utility.
  • [PROMPT_INJECTION]: The skill generates system prompts for subagents using user-supplied text. Ingestion point: Role definitions (scope, constraints) provided by the user in SKILL.md. Boundary markers: None; user strings are interpolated directly into the assets/subagent-cursor-template.md template. Capability: File writing to .cursor/agents/ and definition of subagent behavioral instructions. Sanitization: None; there is no validation to prevent malicious instructions from being included in the generated subagents.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 11, 2026, 05:46 AM