The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests data from untrusted external media to perform tasks such as Automatic Speech Recognition (ASR), Optical Character Recognition (OCR), and script restoration. The extracted text is returned to the agent's context, providing a potential vector for indirect prompt injection. * Ingestion points: Processes are located in scripts such as asr_speech_to_text.py, ocr_text_extract.py, drama_script.py, and video_translation.py. * Boundary markers: No delimiters or explicit instructions to ignore embedded commands are implemented for the processed text outputs returned to the agent. * Capability inventory: The skill possesses network access to Volcengine VOD services and file system access restricted to workspace/, userdata/, and /tmp. * Sanitization: No sanitization or content filtering is applied to the text extracted from media sources before it is processed by the agent.

byted-vod-process-tools