autonomous-orchestration-manager
Pass
Audited by Gen Agent Trust Hub on Sep 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill documentation includes examples for executing a local Python script
scripts/llmops_manager.pyto manage multi-agent workflows, task submission, and monitoring.\n- [INDIRECT_PROMPT_INJECTION]: The task orchestration workflow ingests untrusted user input through the--descriptionfield in thesubmit-taskcommand. This content is likely processed by LLMs to decompose workflows, creating a risk for indirect prompt injection.\n - Ingestion points: User-provided input via the
--descriptionparameter inscripts/llmops_manager.py.\n - Boundary markers: Absent. No specific delimiters or instructions to ignore embedded commands are documented for the orchestration prompt.\n
- Capability inventory: The orchestration script has capabilities for task decomposition, resource allocation, and workflow execution across multiple agents.\n
- Sanitization: Absent. The instructions do not define validation or sanitization requirements for the task descriptions.
Audit Metadata