batch-grill-me
Pass
Audited by Gen Agent Trust Hub on Aug 16, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill explicitly directs the agent to 'dispatch a sub-agent' to retrieve facts from the environment, including the filesystem and system tools. This behavior can be exploited if the agent ingests malicious instructions hidden within local files or tool outputs.
- Ingestion points: Environment and filesystem data are automatically ingested when the agent identifies a need for factual information (
SKILL.md). - Boundary markers: The instructions lack specific boundary markers or delimiters to separate retrieved external content from the agent's core instructions.
- Capability inventory: The skill utilizes sub-agents with access to tools and the local filesystem to populate the 'design tree'.
- Sanitization: No sanitization, filtering, or validation steps are defined for the data retrieved from the environment before it is incorporated into the session context.
Audit Metadata