CREATING_AGENTS

Creating Station Agents

Station agents are intelligent AI assistants that can execute tasks using MCP (Model Context Protocol) tools. This guide covers creating, configuring, and deploying agents.

Agent Fundamentals

What is a Station Agent?

A Station agent consists of:

• Dotprompt Configuration: YAML frontmatter with metadata and input schema
• Multi-role Prompt Structure: System and user role definitions
• Tool Access: MCP tools available to the agent
• Execution Context: Environment and variable configuration

Agent File Structure

Agents are defined in .prompt files with this structure:

---
metadata:
  name: "File Analyzer Agent"
  description: "Analyzes files and directories for insights"
  tags: ["filesystem", "analysis"]
model: gpt-4o-mini
max_steps: 8
tools:
  - "__read_text_file"
  - "__list_directory"
  - "__get_file_info"
  - "__search_files"
input:
  schema:
    type: object
    properties:
      userInput:
        type: string
        description: User input for the agent
      file_path:
        type: string
        description: Path to file or directory to analyze
      analysis_type:
        type: string
        enum: ["basic", "detailed", "security"]
        default: "basic"
    required:
      - userInput
---

{{role "system"}}
You are an expert file analysis agent. When given a file or directory path,
you analyze its contents and provide insights about:
- File types and structure
- Content summaries  
- Potential issues or improvements
- Organization suggestions

Use the provided tools to examine files and directories thoroughly.
Be thorough but concise in your analysis.

{{role "user"}}
{{userInput}}

**Analysis Target:** {{file_path}}
**Analysis Type:** {{analysis_type}}

Creating Agents

Method 1: Interactive CLI Creation

# Start interactive agent creation
stn agent create

# Follow prompts for:
# - Agent name and description
# - Environment selection
# - Tool selection
# - Prompt definition