vigia-method
Pass
Audited by Gen Agent Trust Hub on Oct 22, 2026
Risk Level: SAFEDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill involves processing untrusted third-party source code, which introduces a surface for indirect prompt injection.\n
- Ingestion points: Source code files, route definitions, and API views in the analyzed project.\n
- Boundary markers: No specific delimiters or safety instructions are defined for processing the ingested code content.\n
- Capability inventory: The skill allows file creation in the
.vigia/directory and command execution vianodeandpython3 -Ifor PoC verification.\n - Sanitization: No explicit sanitization or content validation for the source code under review is mentioned.\n- [DYNAMIC_EXECUTION]: The agent is instructed to write and run scripts to confirm vulnerability findings.\n
- Evidence: The 'Prove' stage requires generating and executing
nodeorpython3scripts. These actions are limited to the.vigia/directory and use isolated execution flags (e.g.,-Ifor Python) as a safety measure.
Audit Metadata