podcast-generation
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external data (titles and summaries from bookmarks or tags) and interpolates them directly into a prompt for audio generation in
references/code-examples.md. - Ingestion points: The
generate_audio_narrativefunction takes bookmark content and custom queries as input. - Boundary markers: The current implementation lacks explicit delimiters or instructions to the model to ignore any embedded commands within the bookmark text.
- Capability inventory: The skill uses the Azure OpenAI Realtime API to generate and stream audio narratives.
- Sanitization: No explicit sanitization or filtering of the source content is performed before it is sent to the LLM.
- [EXTERNAL_DOWNLOADS]: The skill incorporates and references scripts and architectural documentation sourced from Microsoft's public repositories, as documented in
UPSTREAM-NOTICE.md. These resources are used to provide the underlying audio conversion and system design logic.
Audit Metadata