Switch the agent's primary model at runtime by patching gateway config.
npx skills add https://github.com/Castrozan/.dotfiles --skill model-switch
Please help me install this Agent Skill. Skill: model-switch Source: https://github.com/Castrozan/.dotfiles/tree/main/agents/skills/model-switch Command: npx skills add https://github.com/Castrozan/.dotfiles --skill model-switch
This skill switches the agent's primary model at runtime by patching the gateway config and letting the gateway auto-restart. Use it when a user requests a model change or when a different model is better suited for the workload.
Use the model-switch skill to change the active model, e.g., "switch to kimi" or "use opus".
High-intent search queries and answers about installing and using this skill.
To switch the agent's primary AI model at runtime, the skill patches the gateway configuration with the new model ID and lets the gateway auto-restart to apply the change. It then confirms the successful model switch to the user.
Yes, you can use model aliases like kimi or opus instead of full model IDs. The skill automatically resolves user-friendly model aliases to their corresponding full model IDs before patching the gateway configuration.
You need to patch the gateway config to change models when a user explicitly requests a model switch or when the system detects a better-suited model for the current workload, such as switching to reduce latency or cost.
Yes, the gateway restarts automatically after applying a model config patch. The skill handles the gateway config patch and relies on this auto-restart mechanism to seamlessly activate the newly selected primary model.
The best way to update the gateway configuration for a new AI model is to use a dedicated model-switch skill that maps aliases to full IDs, patches the config, and triggers an automatic gateway restart to confirm the change.
Switching the active model during a runtime workload allows you to optimize for changing requirements, such as reducing latency or lowering costs by moving from a heavy model like opus to a lighter one like kimi.
JDL440Forward only explicit local-provider model ids and record requestedModel and effectiveModel.
dgethingsMonitor AWS network health via CloudWatch metrics, alarms, and logs.
Sir-chawakornDiagnose runtime performance issues by profiling hotspots and analyzing before-and-after measurements.
dralgorhythmAutomate application deployment to Railway PaaS via CLI and configuration.
