twinmind-security-basics
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill integrates with external LLMs (GPT-4, Claude, Gemini) to process transcription data, creating a surface for indirect prompt injection if malicious content is embedded in transcribed meetings.
- Ingestion points: Meeting audio processed into transcription data in
SKILL.md. - Boundary markers: No explicit instruction delimiters or boundary markers are defined for the AI summarization step.
- Capability inventory: Network health checks via
curland environment variable access (TWINMIND_API_KEY) defined inSKILL.md. - Sanitization: The implementation guides include PII redaction patterns for SSNs, emails, and credit cards in
references/implementation.md. - [SAFE]: The skill demonstrates and recommends robust security practices including timing-safe webhook signature verification, AES-256-GCM encryption for transcripts, and the use of Google Cloud Secret Manager for API key storage.
- [DATA_EXPOSURE]: The skill uses industry-standard patterns for managing credentials via environment variables and demonstrates correct
.gitignoreusage to prevent accidental exposure of.envor.pemfiles.
Audit Metadata