hermes-dojo

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.85). At runtime, scripts/monitor.py reads ~/.hermes/state.db and ingests outsider-authored free text from the messages table (e.g., role='user' content and role='tool' content containing error strings) into its analysis JSON/dashboard, which is then used downstream by the agent/LLM for recommendations and skill patching.

MEDIUM W021: Hidden or invisible Unicode characters detected (potential obfuscation or prompt injection).

• Hidden Unicode characters detected (1 type(s) found)