opencode-orchestrator-creator

Pass

Audited by Gen Agent Trust Hub on Mar 28, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill creates a workspace where the agent is instructed to use curl and jq for network discovery and interaction. It includes a configuration file (.opencode/opencode.json) that pre-authorizes these tools within the agent's sandbox environment.
  • [INDIRECT_PROMPT_INJECTION]: The orchestrator's design involves fetching and processing data from local network endpoints (ports 3001-3010), which represents a potential attack surface if those local services are compromised. * Ingestion points: Server discovery and agent response handling in the generated AGENTS.md file. * Boundary markers: Not present; the logic interpolates raw server responses into messages. * Capability inventory: The orchestrator can execute shell commands and route communications to other agents. * Sanitization: None; data is processed via jq but not validated for natural language injection.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 28, 2026, 09:35 PM
Security Audit — agent-trust-hub — opencode-orchestrator-creator