grilling

Pass

Audited by Gen Agent Trust Hub on Sep 4, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process user-provided content (plans, decisions, or ideas) to build a structured design tree, creating an attack surface for indirect instructions.
  • Ingestion points: User input provided during the "grilling" session to describe the plan or decision in SKILL.md.
  • Boundary markers: The instructions do not define specific delimiters or instructions to disregard embedded commands within the user's plan.
  • Capability inventory: The skill explicitly instructs the agent to dispatch sub-agents to query the environment, including the filesystem and tools, to find facts related to the user's input in SKILL.md.
  • Sanitization: There are no instructions provided to sanitize, escape, or validate the user's input before using it to drive the logic for tool execution or question generation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 4, 2026, 06:52 AM
Security Audit — agent-trust-hub — grilling