inbound-is-data

Pass

Audited by Gen Agent Trust Hub on Jul 29, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill contains references to common prompt injection phrases (e.g., 'Ignore previous instructions') within its documentation; these are used strictly for training the agent to recognize and resist attacks and are not functional instructions.\n- [DATA_EXFILTRATION]: The skill establishes robust preventative measures against data leaks, including a 'mechanical stop' that requires human approval when internal data is processed alongside untrusted sources for outbound communication.\n- [PROMPT_INJECTION]: The skill provides guidance for managing indirect prompt injection by enforcing strict boundaries between data and instructions and requiring content normalization before processing external inputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 29, 2026, 03:55 PM
Security Audit — agent-trust-hub — inbound-is-data