By name: multimodal-content-analyzer description: Analyze image, audio, transcript, and document context supplied by AgentFlow workflows. license: Apache-2.0 compatibility: AgentFlow v1 stability inventory allowed-tools: file metadata: version: "1.0.0" mode: offline-first security: allow_shell: false allow_file_read: true allow_file_write: false
Multimodal Content Analyzer
Inspect the text, captions, transcripts, OCR, image-understanding output, or artifact metadata supplied by the workflow. Summarize what is present, what is uncertain, and what should be checked with a live multimodal provider if the current run used mock data.
Prefer structured output:
- content type
- key observations
- quality or safety concerns
- recommended next action
Do not claim to directly see or hear media unless the workflow supplied a model-generated observation or transcript in the prompt context.