Harness Agents reference

Harness Agents Reference

Worker Agent form field reference

The following fields define a Worker Agent. Required fields are marked in the Required column.

Field	Required	Description	Example
Name	Yes	Human-readable identifier displayed in the catalog and pipeline step picker.	PR Reviewer Agent
Description	No	Free-text summary of what the agent does. Helps teams discover and reuse agents from the catalog.	Reviews PRs for security, schema, and architectural issues.
Instructions	Yes	The system prompt sent to the model at runtime. Supports Harness variable expressions for dynamic context injection. Go to Configure instructions and Harness expressions to review dynamic context injection.	Example below
Model Connector	Yes	The LLM provider connector. When you configure the connector, you select a default model. Go to Configure Model Connectors for supported providers and models.	anthropic_bedrock_99cf4be5
Model Name	No	Optional override for the model used at runtime. If not specified, the agent uses the default model configured on the Model Connector. Accepts an AWS Bedrock inference profile ARN for Anthropic connectors.	arn:aws:bedrock:us-east-1:123456789012:application-inference-profile/a1b2c3d4e5f6
MCP Connectors	No	One or more MCP server connectors granting the agent access to Harness platform data and external services (such as GitHub). Each connector requires a URL and API key.	harness_hosted_mcp
Inputs	No	Named parameters the agent accepts at runtime. Populated from pipeline step outputs, triggers, or manual values. Injected into the agent prompt as context.	llmConnector, modelName, mcpConnectors
Environment Variables	No	Key-value pairs passed to the agent runtime. Used for third-party authentication or model behavior configuration. Supports fixed values or Harness secret expressions.	PLUGIN_HARNESS_CONNECTOR, ANTHROPIC_MODEL

---

Agent definition YAML reference

The following YAML shows the full structure of a Worker Agent definition. This is the YAML visible in the YAML tab when you view or edit an agent in the Worker Agent Catalog (AI > Worker Agents > select agent > YAML tab).

version: 1
agent:
  step:
    group:
      steps:
        - name: Agent
          if: <+Always>
          id: agent
          run:
            container:
              image: pkg.harness.io/vrvdt5ius7uwygso8s0bia/harness-agents/harness-ai-agent:latest
            env:
              PLUGIN_MAX_TURNS: 150
              PLUGIN_TASK: |
                # Agent instructions (system prompt) go here.
                # Supports Harness expressions for dynamic context injection.
              PLUGIN_HARNESS_CONNECTOR: ${{inputs.llmConnector.id}}
              PLUGIN_MCP_FORMAT: harness
              PLUGIN_MCP_SERVERS: <+connectorInputs.resolveList(<+inputs.mcpConnectors>)>
  inputs:
    llmConnector:
      type: connector
      required: true
      default: connector_Anthropic_112e
      ui:
        connectorCategories:
          - AI
    mcpConnectors:
      type: array
      default:
        - connector_Mcp_66c8
      ui:
        component: array
        input:
          inputType: connector
          inputConfig:
            connectorTypes:
              - Mcp
  layout:
    - title: Agent Configuration
      items:
        - llmConnector
        - mcpConnectors

YAML field reference

Field	Description
version	Schema version. Always 1.
agent.step.group.steps	Array of steps executed inside the agent container. The primary step has id: agent and if: <+Always>.
run.container.image	Docker image for the agent runtime. You can override this with your own registry path if you have pulled the Harness base image and published it to your own repository (for example, to add custom tools or dependencies on top of the base image).
env.PLUGIN_MAX_TURNS	Maximum reasoning turns the agent can take per execution.
env.PLUGIN_TASK	The agent's system prompt (Instructions). Supports Harness expressions and built-in Harness environment variables ($HARNESS_ACCOUNT_ID, $HARNESS_ORG_ID, $HARNESS_PROJECT_ID).
env.PLUGIN_HARNESS_CONNECTOR	References the LLM connector input using ${{inputs.llmConnector.id}}.
env.PLUGIN_MCP_FORMAT	MCP protocol format. Use harness for Harness MCP connectors.
env.PLUGIN_MCP_SERVERS	Resolves MCP connector references at runtime using <+connectorInputs.resolveList(...)>.
inputs	Typed parameters the agent accepts. Each input has a type, optional required, default, and ui configuration.
inputs.<name>.ui	Controls how the input renders in the Visual editor. connectorCategories filters connector types; inputConfig.connectorTypes restricts to specific connector kinds.
layout	Defines the Visual editor layout. Groups inputs under titled sections.