hermes-desktop-plugins

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill facilitates the creation of plugins that ingest untrusted data from the Hermes gateway via host.request and host.onEvent, creating a surface for indirect prompt injection.
  • Ingestion points: Data returned by host.request and event payloads from host.onEvent in SKILL.md and templates/plugin.js.
  • Boundary markers: Absent. No instructions are provided to the agent or user regarding the use of delimiters for processed data.
  • Capability inventory: Generated plugins have the capability to register UI components, send notifications, and execute gateway commands.
  • Sanitization: Absent. No sanitization or validation steps are included in the plugin creation procedure or template.
  • [COMMAND_EXECUTION]: The skill generates and writes JavaScript files to the local filesystem ($HERMES_HOME/desktop-plugins/) which are subsequently executed by the host application. This behavior constitutes dynamic execution and a persistence mechanism.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 01:33 AM
Security Audit — agent-trust-hub — hermes-desktop-plugins