Andruia Consultant
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill analyzes pre-existing files in the workspace to perform technical scans and prescribe improvements. This creates a surface where malicious instructions embedded in project files could be processed by the agent.
- Ingestion points: Local project files, including 'src' and 'package.json'.
- Boundary markers: No explicit delimiters or warnings are defined to separate ingested data from instructions.
- Capability inventory: The skill reads workspace files and writes markdown roadmaps (tareas.md, plan_implementacion.md).
- Sanitization: No sanitization or validation of the ingested file content is specified.
Audit Metadata