grilling

Pass

Audited by Gen Agent Trust Hub on Sep 22, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided plans, decisions, or ideas which serves as an ingestion point for untrusted data. This creates a surface where instructions embedded in the user's plan could be followed by the agent.
  • Ingestion points: User input regarding plans or ideas processed in SKILL.md.
  • Boundary markers: There are no explicit instructions, delimiters, or system prompts to isolate user input or to warn the agent against executing instructions potentially contained within the user-provided plans.
  • Capability inventory: The instructions explicitly authorize the agent to use sub-agents to access the filesystem and other environment tools to find facts (SKILL.md).
  • Sanitization: The skill does not implement sanitization, validation, or filtering of the user-provided content before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 22, 2026, 11:39 PM
Security Audit — agent-trust-hub — grilling