iflytek-ocr-invoice
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (images and PDFs) provided by the user to extract text via OCR. This extracted text is then returned to the agent's context without sanitization or instruction-boundary delimiters. An attacker could provide a document containing text instructions designed to manipulate the agent's behavior after the OCR result is parsed.
- Ingestion points: The
scripts/invoice.pyscript reads files from a user-supplied filesystem path. - Boundary markers: The script provides structured output but does not include explicit warnings to the agent to ignore any natural language instructions found within the OCR results.
- Capability inventory: The script possesses file-read capabilities and performs HTTPS POST requests to the iFlytek API endpoint.
- Sanitization: No filtering or sanitization is performed on the text returned by the OCR engine before it is passed back to the agent.
- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill follows standard security practices by requiring sensitive API credentials (
XFYUN_APP_ID,XFYUN_API_KEY,XFYUN_API_SECRET) to be managed via environment variables rather than hardcoding them. It transmits image data to the iFlytek API for processing as part of its primary function.
Audit Metadata