zj-grilling
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to gather facts from the environment, including the filesystem and external tools, to answer its own questions. This data ingestion path creates a surface for indirect prompt injection, where content within files or tool outputs could provide instructions that override the agent's intended behavior.
- Ingestion points: Filesystem and external tools via sub-agents (SKILL.md).
- Boundary markers: None identified; the skill does not explicitly instruct the agent to ignore instructions embedded in the gathered data.
- Capability inventory: The skill has the capability to read files and invoke tools.
- Sanitization: No sanitization or validation of the environmental facts is mentioned.
- [COMMAND_EXECUTION]: The instructions explicitly direct the agent to "dispatch a sub-agent to find" facts from tools and the filesystem. This implies the execution of shell commands or tool invocations to explore the local environment without direct user intervention for each step.
Audit Metadata