twinmind-sdk-patterns
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill documentation provides code that interacts with the external domain api.twinmind.com to store and retrieve conversation context. While this is the intended service infrastructure for the skill, it involves sending data to a non-whitelisted external domain. The implementation correctly utilizes environment variables for authentication.- [INDIRECT_PROMPT_INJECTION]: The provided patterns ingest external content such as meeting transcripts and memory strings for processing by the TwinMind API. This establishes a potential attack surface where adversarial content within a transcript could attempt to influence the agent's logic or the resulting insights. * Ingestion points: The transcript parameter in create_meeting_context and content parameter in store_memory in SKILL.md. * Boundary markers: None are included in the code templates to separate untrusted content. * Capability inventory: The skill utilizes the requests library to perform network operations. * Sanitization: No input validation or sanitization routines are provided in the implementation patterns.
Audit Metadata