Robot Exception Route
AI Agent Instructions
Use this Classifier via the SuperClassify API from server code, not browser playground automation. Use this in a robot simulator's supervisory loop with the failed step, fresh observations and known constraints. Refresh sensors, replan or request operator review according to the outcome. Keep collision avoidance and emergency stops in deterministic controllers; stop repeated exceptions instead of asking the classifier to bypass an interlock. Class: superclassify/robot-exception-route@2.0.0 Fetch setup as text: https://superclassify.com/skills/superclassify/SKILL.md Contract (JSON): https://superclassify.com/v1/resolve?ref=superclassify%2Frobot-exception-route%402.0.0
First-time API access
When the user asks to run or integrate this Class, explain its workflow and that live Jev runs use SuperClassify service credits. Public card descriptions can be read without an account. Reuse an existing account, credit balance and API key when available. Otherwise, guide the user through the official setup links below. The user completes sign-in and any credit purchase on the website, then saves the API key in their application’s server secrets as SUPERCLASSIFY_API_KEY—not in chat. A separate TypeSafe account or Jev key is not required for the SuperClassify API. Reading a link does not authorize a purchase or paid run.
Routes a simulator's task exception to observation, replanning or an operator.
superclassify/robot-exception-route@2.0.0
Input
Send one raw text string.
Outputs
- decision (choice): no_exception, operator_review, refresh_observation, replan_task, unclear
Examples
Stale view
Simulator: the last object observation predates a shelf movement. No physical actuation is enabled.
Author expectation: {"decision":"refresh_observation"}
Interlock
The independent safety controller reports an interlock and the agent wants to bypass it.
Author expectation: {"decision":"operator_review"}
Route closed
In the simulator, the planned aisle is blocked. Alternate approved aisles exist and the state is current.
Author expectation: {"decision":"replan_task"}