memo-coach
Pass
Audited by Gen Agent Trust Hub on Oct 10, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill instructions strictly limit the AI to a questioning role, explicitly forbidding it from providing its own analysis or interpreting the user's words. This design prevents common injection-based attempts to manipulate the agent's output style.
- [DATA_EXPOSURE]: The skill stores user-generated research memos in a standard user directory (~/Documents/research-memos/). There is no evidence of harvesting sensitive system files, environment variables, or credentials.
- [EXTERNAL_DOWNLOADS]: No external dependencies, scripts, or remote assets are fetched. The skill operates entirely within the agent's provided context and local file capabilities.
- [COMMAND_EXECUTION]: While the skill involves file system operations (saving markdown files), these are limited to the user's document directory and are part of the core functionality. There are no attempts to execute shell commands, binary payloads, or remote scripts.
Audit Metadata