disable-model-invocation

To prevent an AI model from automatically invoking a skill, you can use a manual invocation control mechanism. This requires adding a specific flag to the skill's configuration file to block autonomous execution and ensure explicit user triggering.

The AI model executes sensitive operations without user confirmation because the skill lacks manual invocation restrictions. Implementing a disable-model-invocation flag ensures the model cannot autonomously trigger actions, requiring deliberate manual or programmatic execution for safety.

The best way to require manual user consent before modifying system configurations is to restrict autonomous AI execution. By marking the skill with a disable-model-invocation flag, you enforce human oversight and prevent the model from performing critical operations directly.

To configure a skill for explicit manual invocation only, you must modify its SKILL.md file. Adding the 'disable-model-invocation: true' flag prevents the AI model from directly calling the skill, ensuring it only runs when deliberately triggered by a user or system.

You need to disable model invocation for a skill when it performs sensitive operations or modifies system configurations. This safety mechanism is necessary to prevent unintended autonomous execution by the AI model and ensure direct human oversight before critical actions occur.