opendia-harness

Pass

Audited by Gen Agent Trust Hub on Sep 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the user to execute shell commands to manage plugins and run a server process.
  • Evidence includes commands like grok plugin install opendia --trust and npx opendia to initialize the harness.
  • [EXTERNAL_DOWNLOADS]: The skill directs the user to download assets from external sources.
  • It references GitHub releases at github.com/aeonfun/opendia/releases for browser extension components.
  • It utilizes npx -y opendia to fetch and run the latest version of the server from the NPM registry.
  • [INDIRECT_PROMPT_INJECTION]: As a setup guide for a browser-integrated MCP harness, the skill establishes a data flow between external web content and the agent.
  • Ingestion points: Browser extension connected via WebSocket (ws://localhost:5555) provides data to the agent.
  • Boundary markers: Not explicitly defined in this configuration skill; handling is deferred to the runtime playbook.
  • Capability inventory: The harness facilitates tool calls and browser interaction within the agent environment.
  • Sanitization: Not specified within the setup instructions; depends on the underlying OpenDia implementation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 12, 2026, 08:00 AM