Installation

$ npx skills add smithery/aws-samples --skill modify-agent-prompts

Summary

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.

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared

Cursor Not declared

Codex Not declared

GitHub Copilot Not declared

Windsurf Not declared

Gemini CLI Not declared

Cline Not declared

OpenCode Not declared

Package contents

Files included with this skill beyond the listing page.

skill md SKILL.md 3,734 B

docs SUMMARY.md 102 B

SKILL.md

Modify Agent Prompts and Configuration

Agent Architecture

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

Common Modifications

1. Modifying Document Review Prompts

File: agent.py

Key functions:

getdocumentreview_prompt() - Main entry point
getdocumentreviewpromptwithcitations() - With citations
getdocumentreviewprompt_legacy() - Without citations

JSON Schema:

{
  "result": "pass" | "fail",
  "confidence": <number 0-1>,
  "explanation": "<detailed reasoning>",
  "shortExplanation": "<max 80 chars>",
  "pageNumber": <integer from 1>,
  "citations": ["<quoted text>", ...]
}

2. Modifying Image Review Prompts

File: agent.py, function getimagereview_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>]
}

3. Model Configuration

Environment Variables (set in CDK):

DOCUMENTPROCESSINGMODELID, IMAGEREVIEWMODELID, BEDROCK_REGION
ENABLECITATIONS, ENABLECODE_INTERPRETER

Change models via CDK:

cdk deploy -c rapid.documentProcessingModelId="global.anthropic.claude-opus-4-6-v1"

Add new model to modelconfig.py in getmodelconfigs():

"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,
)

4. Adding New Environment Variables

Define in cdk/lib/parameter-schema.ts
Pass to Lambda in cdk/lib/constructs/agent.ts
Read in agent.py with os.environ.get()

5. Adding New Tools

For detailed tool creation guide, see [references/TOOL-CREATION.md](references/TOOL-CREATION.md).

Quick Reference

Modification	Location	Search For
Document prompt schema	agent.py	`getdocumentreviewprompt`
Image prompt schema	agent.py	`getimagereview_prompt`
Confidence guidelines	agent.py	`confidence_guidelines`
Tool usage instructions	agent.py	`buildtoolusagesection`
Model capabilities	model_config.py	`getmodel_configs`
Environment variables	CDK agent.ts	`environment:`

Troubleshooting

Issue	Check
Citations not working	`ENABLECITATIONS` env var, model supports citations in `model``config.py`
Model not found	Model ID format, available in `BEDROCKREGION`, 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

After Modification

Run /build-and-format if Python code changed
Run /deploy-cdk-stack if CDK changes made
Test with sample documents and monitor CloudWatch logs

modify-agent-prompts