twinmind-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Apr 4, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted transcript data to automate task creation and email generation, creating an indirect prompt injection surface.\n
- Ingestion points: Transcript data is processed via the ActionItemExtractor class in references/implementation.md based on transcript IDs.\n
- Boundary markers: No delimiters or instructions were found to isolate transcript content from agent instructions.\n
- Capability inventory: Extracted data is used to create tasks in Asana/Linear and generate emails via TwinMind.\n
- Sanitization: The code does not implement explicit sanitization for the transcript content before it is used in automated task creation or email generation.\n- [SAFE]: The skill connects to well-known project management services (Asana, Linear) and the TwinMind API for core workflow automation. These network operations are consistent with the skill's purpose and use environment variables for secure credential management.
Audit Metadata