agentic-ai-engineering/course/labs/L4-agent-chain/starter.yaml

60 lines
1.8 KiB
YAML

# Lab 4.12: Agent Chain Pipeline
#
# Define a plan → build → review → verify pipeline.
# Fill in the prompt templates for each step.
#
# Variables:
# $INPUT = output of previous step (or user prompt for first step)
# $ORIGINAL = the user's original prompt
steps:
- agent: planner
prompt: |
You are a senior software architect.
Create a detailed implementation plan for: $INPUT
Your plan should include:
- Architecture overview
- Step-by-step implementation tasks
- Files to create/modify
- Key design decisions
Format your plan with clear sections and task breakdowns.
- agent: builder
prompt: |
You are a senior software engineer.
Implement the following plan: $INPUT
Rules:
- Write production-quality code
- Include error handling
- Add comments for complex logic
- Run tests after implementation
# TODO: What specific instructions should the builder follow?
- agent: reviewer
prompt: |
You are a senior code reviewer.
Review this implementation: $INPUT
Check for:
- Correctness (does it work?)
- Security (are there vulnerabilities?)
- Performance (are there bottlenecks?)
- Style (does it follow conventions?)
# TODO: What review criteria should be added?
- agent: verifier
prompt: |
You are a verification agent with READ-ONLY access.
Verify the following claims from the review: $INPUT
For each claim, check actual file contents and report:
- PERFECT: All claims verified, no gaps
- VERIFIED: All claims pass, minor gaps
- FEEDBACK: Some claims failed, corrections needed
- FAILED: Cannot verify, escalation required