orchestrate-agents
Pass
Audited by Gen Agent Trust Hub on Jul 22, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
cursor-agent --list-modelscommand to check for available model IDs and prevent errors during dispatch. - [EXTERNAL_DOWNLOADS]: The instructions provide links to external GitHub repositories for required plugins, such as
github.com/codemeall/grok-plugin-cc, which are associated with the skill's author. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it processes external data from sub-agents and user arguments. Ingestion points: User input via
$ARGUMENTSand results returned by worker agents. Boundary markers: Employs a specificresult contractto format worker outputs. Capability inventory: Orchestrates plugins with file-writing capabilities and executes local CLI tools. Sanitization: Uses model ID validation and alias mapping to restrict worker selection.
Audit Metadata