datasecurer
Pass
Audited by Gen Agent Trust Hub on Sep 6, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data from project briefs and architecture documents. Ingestion points: Files located at ideator-output/project-brief.md and constructor-output/architecture-doc.md. Boundary markers: None identified. Capability inventory: Limited to read, write, and directory creation. Sanitization: No specific sanitization of external text is performed. The skill uses this data for documentation purposes and does not interpret it as commands.
- [COMMAND_EXECUTION]: The skill uses the exec tool to run a hardcoded mkdir command for output management. This operation is localized to the datasecurer-output directory and does not incorporate unsanitized user input.
Audit Metadata