n8n-agents
Pass
Audited by Gen Agent Trust Hub on Oct 24, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: This skill defines an architecture for AI agents that ingest untrusted data from external chat surfaces (Slack, Discord, Telegram) and webhooks, as documented in
references/CHAT_AGENT_PATTERNS.mdandreferences/TOOLS.md. These agents have access to significant capabilities, including database writes (Notion) and sub-workflow execution (references/EXAMPLES.md). However, the skill provides robust security patterns to mitigate injection risks, such as using human-in-the-loop approval nodes (references/HUMAN_REVIEW.md) for destructive tool calls and using deterministic parameter plumbing (references/TOOLS.md) to ensure sensitive identifiers like user IDs and session keys are handled by the system rather than the AI model.
Audit Metadata