submittal-intelligence
Pass
Audited by Gen Agent Trust Hub on Mar 17, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it processes untrusted external data from submittal documents (PDFs, images, and data sheets) to populate compliance matrices and generate review comments. Maliciously crafted content within these documents could attempt to override the agent's logic to force an 'Approved' status or inject deceptive information into project communications.
- Ingestion points: Submittal documents and metadata parsed in Workflow Steps 1, 4, and 5a.
- Boundary markers: The workflow does not explicitly mention the use of delimiters or instructions to the model to ignore potential commands embedded within the extracted text.
- Capability inventory: The skill is capable of updating the submittal-log.json file, generating project-wide alerts via /morning-brief, and drafting comments for formal RFI responses.
- Sanitization: No explicit sanitization, validation, or schema-checking logic is described for the technical data extracted from external documents before it is used to determine compliance status.
Audit Metadata