The Agent Skills Directory

[SAFE]: The skill instructions focus on a structured interviewing process for design review. No evidence of prompt injection, data exfiltration, or unauthorized command execution was found.
[INDIRECT_PROMPT_INJECTION]: The skill has an inherent surface for indirect prompt injection as it processes untrusted data from the user and the codebase.
Ingestion points: User-provided plans/designs and the local codebase (SKILL.md).
Boundary markers: None specified in the prompt.
Capability inventory: Uses the AskUserQuestion tool and performs codebase exploration (file reads).
Sanitization: No specific sanitization or filtering is mentioned for the ingested content.
Note: While the surface exists, the risk is minimal as the skill does not utilize high-privilege tools or network operations.

grill-me