[object Object]

{Agent Name}

Purpose

{One sentence describing what this agent does and what value it provides.}

Context

Agent receives:

• task: Task description from orchestrator
• input_path: Path to input data file
• output_path: Where to write results
• session_id: Current session identifier

Additional context (if needed):

• config: Relevant configuration from L0/L1
• constraints: Operational constraints

Instructions

1. Read Input

   • Load data from {input_path}
   • Validate against expected schema

2. Process

   • {Step-by-step processing logic}
   • {Each step should be atomic and verifiable}

3. Validate Output

   • Check output meets quality criteria
   • Ensure all required fields present

4. Write Output

   • Write to {output_path} using io-yaml-safe
   • Include metadata (timestamp, source references)

Constraints

• {Constraint 1: e.g., "Do not make external API calls beyond provided tools"}
• {Constraint 2: e.g., "Maximum processing time: 5 minutes"}
• {Constraint 3: e.g., "Do not modify input files"}

Output Schema

# {output_file}.yaml
metadata:
  agent: {agent-name}
  created_at: {timestamp}
  input_source: {input_path}

results:
  # {Define your output structure}
  items: []
  summary: ""

quality:
  items_count: 0
  validation_passed: true

Quality Criteria

• All required fields populated
• No hallucinated data (everything traceable to input)
• Output validates against schema
• Processing completed without errors

Examples

Input

# Example input structure
task: "Process items"
items:
  - id: 1
    data: "..."

Output

# Example output structure
metadata:
  agent: {agent-name}
  created_at: "2026-01-30T10:00:00Z"

results:
  items:
    - id: 1
      processed: true
      result: "..."
  summary: "Processed 1 item successfully"