grill-me
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to resolve facts through local files, documentation, and tools in the available environment to automate project understanding. This capability creates an attack surface where malicious or confusing instructions embedded in project files could influence the agent's logic during the interview or be included in the final brief.
- Ingestion points: Project files, documentation, and environment-available tools (SKILL.md, Instruction 1).
- Boundary markers: No explicit markers or delimiters are defined to isolate data read from files from the skill's core instructions.
- Capability inventory: The skill performs interviews, summarization, and brief generation, intended as inputs for development tools such as create-software, create-internal-app, and create-landing.
- Sanitization: No automated sanitization is specified for ingested file content; however, the skill mandates a human-in-the-loop safeguard by requiring explicit user confirmation of the final brief before any actions are taken (SKILL.md, Instruction 9).
Audit Metadata