silicon-paddle-ocr
Pass
Audited by Gen Agent Trust Hub on Sep 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests images which may contain text intended to override agent instructions, potentially leading to indirect prompt injection. \n
- Ingestion points: The
scripts/ocr_skill.pyscript reads image files provided by the user. \n - Boundary markers: Absent. No delimiters or specific instructions are included in the prompt to ensure the model ignores instructions found within the image data. \n
- Capability inventory: The skill communicates with the SiliconFlow API and can output results to the terminal or local files; it does not perform automated actions or code execution based on the extracted text. \n
- Sanitization: None. The text content recognized from images is returned or saved without validation.
Audit Metadata