lindy-data-handling
Pass
Audited by Gen Agent Trust Hub on Sep 9, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: Documentation of untrusted data ingestion points (customer data, web crawls) and associated agent capabilities (Read, Write, Edit). The skill provides mitigation strategies including prompt boundary rules and automated PII redaction logic.\n
- Ingestion points: Tasks, Memory, Knowledge Base, and Integrations described in SKILL.md.\n
- Boundary markers: Recommended 'Data Handling Rules' block for agent prompts to guide redaction and escalation.\n
- Capability inventory: Tools for file system access and SDK-based agent execution.\n
- Sanitization: Implementation of regex-based redaction provided in references/implementation-guide.md.\n- [EXTERNAL_DOWNLOADS]: Code samples in references/implementation-guide.md reference the
@lindy-ai/sdkpackage for interaction with the vendor platform.
Audit Metadata