Skip to main content

Advanced Skills Configurations

Overview​

Beyond adding a skill to an agent, Serenity* Star lets you control how skills behave at execution time. These advanced configurations do not change the agent's saved design; they shape a single request.

  • Include & exclude skills — narrow the set of skills available for one execution with a per-request allow-list or deny-list, without editing the agent.
  • Input parameter mapping — pass values in the execution request, map them to your agent's input parameters, and have them resolve as Liquid variables inside a skill's configuration.
  • AI-generated skill parameters — declare parameters the agent completes itself when no value exists in the execution context, and write the descriptions it follows to fill them.
  • HTTP request bodies & content types — choose a body type and describe a request body for JSON, plain text, XML, URL-encoded and multipart endpoints.

The pages below cover each one in detail, with request examples.

Include & Exclude Skills

Restrict which skills an agent may use for a single execution with the reserved skills parameter — an include allow-list or an exclude deny-list.

Input Parameter Mapping

Send values in the agent request, map them to input parameters, and reference them as Liquid variables inside a skill's configuration.

AI-Generated Skill Parameters

Let the agent complete a skill's parameters when the execution context has no value for them, and write the descriptions it follows to fill them.

HTTP Request Bodies & Content Types

How to describe a request body for each supported content type: JSON, plain text, XML, URL-encoded forms and multipart forms.

note

These configurations apply to skills you have already added to an agent. To add a skill in the first place, see Introduction; for the full catalogue, see Available Skills.