Skills

Configure LLM-backed micro-agent capabilities with system prompts, schemas, and tool access.

Skills are companion agent capabilities. They are LLM-backed micro-agents that execute specific tasks within a conversation. Unlike Classic API tools (which run versioned code packages), skills are configured declaratively with a system prompt, input/output schema, and model selection.

Different from Classic API Tools. Classic API Tools are versioned code packages from Git repos. Platform API Skills are LLM-backed declarative micro-agents with prompt-based configuration. Choose Tools for custom code execution, Skills for LLM-native reasoning.

API Surface

Use the Skills API reference for the current operation list, request/response schemas, filters, pagination parameters, and reference lookup route. This page covers modeling guidance for skill configuration.

Core Fields

Field	Type	Description
id	string	Unique identifier
slug	string	URL-friendly identifier (lowercase, hyphens/underscores, 2-63 chars). Immutable after creation.
name	string	Human-readable name
description	string	What the skill does
system_prompt	string or null	Instructions for the skill's LLM
input_schema	object	JSON Schema defining expected inputs
result_schema	object or null	JSON Schema defining structured outputs
enabled	boolean	Whether the skill is available for execution (default true)

Input Schema Compatibility

A skill input_schema is used as an LLM tool schema during execution. Keep it in the supported tool-schema subset:

• The root schema must describe a single JSON object.
• Do not use a top-level $ref, anyOf, oneOf, or allOf.
• Array-typed properties must declare items.

Unsupported shapes are rejected on create/update with 422 because they can cause the LLM tool call to fail silently at runtime.

Execution Configuration

Field	Type	Default	Description
model	string	claude-sonnet-4-6	LLM model for skill execution. Must be an Anthropic model ID (claude*/anthropic*). Contact your Amigo representative for available models.
max_tokens	integer	4096	Maximum output tokens
max_result_chars	integer	2000	Maximum result text length
max_input_tokens	integer or null	null	Maximum input tokens per execution (1 to 2,000,000)
timeout_s	float	60.0	Execution timeout in seconds (0.1-900)
temperature	number	Optional	Sampling temperature (0-1). Controls response randomness. Leave unset (null) to use the model default. Model-gated - models that do not support sampling parameters ignore this field silently. Set at most one of temperature or top_p; if both are provided, the runtime keeps temperature and drops top_p.
top_p	number	Optional	Nucleus sampling (0-1). Controls response diversity. Leave unset (null) to use the model default. Model-gated - models that do not support sampling parameters ignore this field silently. Set at most one of temperature or top_p.
thinking_effort	string or null	null	low, medium, or high
enable_caching	boolean	true	Cache LLM responses
enable_citations	boolean	false	Include source citations
use_structured_output	boolean	false	Force structured JSON output
max_agent_turns	integer	20	Maximum tool-use turns per execution (1-200)
checkpoint_enabled	boolean	true	Enable checkpointing
approval_required	boolean	false	Require human approval before execution

Skills execute as multi-turn LLM agents: the model can call the skill's configured tools across up to max_agent_turns turns before producing its final result.

Integration with External APIs

Skills can call external APIs through Integrations and inline tool definitions:

Field	Type	Description
integration_tools	array	References to integration endpoints: { integration, endpoint } (where integration is the integration name and endpoint is the endpoint name within that integration)
static_tools	array	Inline tool definitions with name, description, and input schema

References

Before modifying or deleting a skill, check what depends on it:

The references operation returns context graphs (with specific states) and services that reference the skill's slug. This prevents breaking live conversation flows by accidentally removing a skill that is in use.

Testing

Use the tool testing endpoints to execute a skill in isolation without starting a conversation. Useful for validating prompt engineering and schema design before wiring into a context graph.

API Reference

• Skills