reasoning-orchestrator
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a structured, multi-step reasoning architecture. It acts as a router that selects appropriate internal tools based on a classification (triage) of the user's input.
- [COMMAND_EXECUTION]: The orchestrator utilizes a 'Task' tool to spawn subagents. The instructions provided to these subagents are strictly limited to applying methodologies from specific, predefined skill files within the 'skills/' directory. This follows the principle of least privilege for the subtasks.
- [DATA_EXPOSURE]: The skill uses the 'view' command to read methodological instructions from 'skills/[skill-name]/SKILL.md'. This is an internal lookup mechanism for the agent's own capabilities and does not expose sensitive system configuration, credentials, or user data.
- [PROMPT_INJECTION]: While the skill uses imperative language and 'CRITICAL' markers, these are used to maintain the integrity of the reasoning process (e.g., preventing the agent from skipping steps or hallucinating reasoning instead of using the provided skills). These instructions do not attempt to bypass safety filters or exfiltrate data.
- [DATA_EXFILTRATION]: No network operations or external data transfer mechanisms are present in the skill's logic.
Audit Metadata