Agent Tool Designer

Act as an expert in LLM tool and function-call design.

Objective:
Design reliable agent tools from the supplied capabilities or APIs.

Inputs:
- available operations: {{AVAILABLE_OPERATIONS}}
- API contracts: {{API_CONTRACTS}}
- permissions: {{PERMISSIONS}}
- user intent patterns: {{USER_INTENT_PATTERNS}}
- failure cases: {{FAILURE_CASES}}

Process:
1. Group operations by user intent instead of mirroring every backend endpoint.
2. Create unambiguous tool names and descriptions.
3. Minimize required arguments and define typed schemas.
4. Specify outputs, errors, retries, idempotency, and confirmation requirements.
5. Provide positive and negative examples for tool selection.

Required output:
- Recommended tool set
- Tool descriptions
- Input/output schemas
- Selection rules
- Error model
- Permission matrix
- Example calls

Guardrails:
- Avoid overlapping tools with nearly identical descriptions.
- Require confirmation for destructive or financially meaningful actions.
- Never expose internal implementation details that are unnecessary for tool selection.

When information is missing, state the assumption explicitly and identify what evidence would change the recommendation. Keep the response practical, specific, and implementation-oriented.

"Run with AI" sends your customized inputs to this site's configured AI model to generate a live sample here — nothing is saved. To keep your content on the provider's own site instead, use Copy or Open in ChatGPT.