ocr
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: Executes the vendor-provided casedev CLI to perform OCR processing, track job status, and retrieve word-level positional data from documents.- [EXTERNAL_DOWNLOADS]: Allows the agent to fetch documents from remote URLs for processing via the --document-url flag.- [PROMPT_INJECTION]: Exposure to indirect prompt injection from the content of processed documents (PDFs and images) being returned to the agent context.
- Ingestion points: External files processed via URLs or local uploads (e.g., ./scanned-contract.pdf).
- Boundary markers: None identified in the CLI command structure.
- Capability inventory: Executes CLI-based subprocesses that perform network requests and file reads.
- Sanitization: The skill returns extracted text directly from the OCR engine without explicit sanitization steps.
Audit Metadata