god-meta-conductor

To prevent AI coding hallucinations and maintain focus, you can load a meta-conductor skill that enforces zero-hallucination discipline and task path locking. It requires the AI to verify facts through external sources rather than memory, ensuring precise, verified technical outputs.

Task path lock in AI coding sessions ensures the AI remains strictly focused on the current task until completion or explicit termination. It treats all mid-task messages as extensions of the original request, preventing scope drift and unrequested actions through strict scope enforcement.

To stop an AI from making assumptions and generating unverified code, enforce a no-assumptions-without-declaration rule. This requires the AI to explicitly declare any assumptions made during technical tasks, promoting source verification over memory and prioritizing completeness over speed.

Yes, AI coding accuracy improves when prioritizing completeness over speed because it enforces a self-correction loop. This approach encourages the AI to review and correct its outputs before delivery, ensuring technical accuracy and reducing the risk of errors in the final code.

You should use strict focus control for technical tasks requiring precise, verified outputs where uncertainty is unacceptable. It is essential when you need zero-hallucination discipline, explicit task completion signaling, and restraint of the AI's thinking model before code generation.

Yes, you can send mid-task messages to an AI without resetting the coding scope because the system treats all messages as extensions unless an explicit stop command is received. This ensures the AI remains locked on the requested scope and performs only explicitly requested actions.