// Modify the review-item-processor agent's prompts, model configuration, confidence thresholds, JSON output schemas, and tool setup. Use when changing document or image review prompts, adjusting model IDs, modifying output format, updating citation settings, or adding new tools to the agent.
| name | modify-agent-prompts |
| description | Modify the review-item-processor agent's prompts, model configuration, confidence thresholds, JSON output schemas, and tool setup. Use when changing document or image review prompts, adjusting model IDs, modifying output format, updating citation settings, or adding new tools to the agent. |
review-item-processor/
├── agent.py # Prompt generators + agent execution
├── model_config.py # Model registry with capabilities
├── tools/
│ ├── factory.py # Dynamic tool creation
│ ├── knowledge_base.py # Bedrock KB integration
│ ├── code_interpreter.py
│ └── mcp_tool.py # MCP tool support
└── index.py # Lambda handler entry point
File: agent.py
Key functions:
get_document_review_prompt() - Main entry point_get_document_review_prompt_with_citations() - With citations_get_document_review_prompt_legacy() - Without citationsJSON Schema:
{
"result": "pass" | "fail",
"confidence": <number 0-1>,
"explanation": "<detailed reasoning>",
"shortExplanation": "<max 80 chars>",
"pageNumber": <integer from 1>,
"citations": ["<quoted text>", ...]
}
File: agent.py, function get_image_review_prompt()
JSON Schema (Nova models add boundingBoxes):
{
"result": "pass" | "fail",
"confidence": <number 0-1>,
"explanation": "<detailed reasoning>",
"shortExplanation": "<max 80 chars>",
"usedImageIndexes": [<list of indexes>]
}
Environment Variables (set in CDK):
DOCUMENT_PROCESSING_MODEL_ID, IMAGE_REVIEW_MODEL_ID, BEDROCK_REGIONENABLE_CITATIONS, ENABLE_CODE_INTERPRETERChange models via CDK:
cdk deploy -c rapid.documentProcessingModelId="global.anthropic.claude-opus-4-5-20251101-v1:0"
Add new model to model_config.py in _get_model_configs():
"model-id": ModelConfig(
model_id="model-id",
supports_document_block=True,
supports_citation=True,
supports_caching=True,
input_per_1k=0.XXX,
output_per_1k=0.XXX,
)
cdk/lib/parameter-schema.tscdk/lib/constructs/agent.tsagent.py with os.environ.get()For detailed tool creation guide, see references/TOOL-CREATION.md.
| Modification | Location | Search For |
|---|---|---|
| Document prompt schema | agent.py | _get_document_review_prompt |
| Image prompt schema | agent.py | get_image_review_prompt |
| Confidence guidelines | agent.py | confidence_guidelines |
| Tool usage instructions | agent.py | _build_tool_usage_section |
| Model capabilities | model_config.py | _get_model_configs |
| Environment variables | CDK agent.ts | environment: |
| Issue | Check |
|---|---|
| Citations not working | ENABLE_CITATIONS env var, model supports citations in model_config.py |
| Model not found | Model ID format, available in BEDROCK_REGION, added to model_config.py |
| Tool not available | Tool configuration in event payload, registered in factory.py |
| Wrong output format | JSON schema in prompt, {language_name} placeholders |
/build-and-format if Python code changed/deploy-cdk-stack if CDK changes madeDeploy RAPID application AWS infrastructure using CDK, including stack deployment, database migrations, and post-deployment verification. Use when explicitly asked to deploy, when running CDK bootstrap for first-time setup, or when deploying after code or infrastructure changes. Only execute when user says "please deploy".
Add new example use cases to the Examples feature, including file setup, metadata registration, thumbnail generation, and link verification. Use when a contributor provides PDF/PNG/TXT files for a new use case, when adding entries to the Examples page, or when verifying existing example GitHub URLs are accessible.
Run build verification and code formatting across all RAPID project components (backend, frontend, CDK). Use after implementing code changes, before committing, after modifying Prisma schema, or before creating a pull request.
Modify CDK Step Functions workflows (ReviewProcessor and ChecklistProcessor) for the RAPID application, including workflow step changes, Map State concurrency, retry/timeout configuration, error handling, and parameter updates. Use when changing workflow step sequences, adjusting concurrency, modifying retry logic, or adding/removing workflow steps.
Create implementation plans for backend API endpoints and frontend features following RAPID layered architecture (domain/usecase/routes pattern, SWR hooks, feature-based organization). Use when adding or modifying backend APIs, frontend components, database schemas, or repository implementations. Also use for refactoring backend/frontend code.
Run repository integration tests against a local MySQL database for the RAPID backend. Use after implementing new repository methods, modifying Prisma database schema, testing use cases that depend on database access, or verifying data access patterns.