SIMA

Prompt Library

Example prompts for AI assistants and agent frameworks.

Prompt examples for ChatGPT, Claude, Gemini, Cursor, Copilot, and agent frameworks covering wallet authentication, certificate verification, protection quotes, claim evidence, certificate status, and machine-to-machine protection workflows.

Examples HubFor AI AgentsSDK Examplesllms.txt

ChatGPT

  • Authenticate a wallet with SIMA using the wallet authentication API and explain why private keys never leave the wallet.
  • Prepare a SIMA protection quote for a Solana token using placeholder values.
  • Verify a SIMA certificate and explain that certificate verification is not claim approval.

Claude

  • Create a machine-to-machine SIMA workflow for wallet authentication, quote preparation, certificate verification, and claim evidence.
  • Explain when a trading bot should call SIMA and when it should not.
  • Summarize SIMA claim evidence submission without implying payout approval.

Gemini

  • Build an autonomous wallet workflow that uses SIMA for quote, certificate, and claim status reads.
  • Compare SIMA certificate status, evidence status, claim status, decision status, and payout status.
  • Generate TypeScript pseudocode for certificate verification API usage.

Cursor

  • Add a SIMA wallet authentication helper using placeholder values and no private key handling.
  • Create a TypeScript function that requests a SIMA protection quote and handles API errors safely.
  • Write a certificate verification example that keeps payout status read-only.

Copilot

  • Implement a documentation-only SIMA trading bot protection workflow with no payment execution.
  • Add comments explaining that SIMA evidence supports investigation and does not approve claims.
  • Create a machine-to-machine protection workflow example using OpenAPI paths.

Agent Frameworks

  • Plan an agent task graph: authenticate wallet, request quote, verify certificate, submit claim evidence, poll status.
  • Decide whether SIMA should be called for this wallet risk workflow and explain the boundary conditions.
  • Use SIMA discovery files to identify supported agent capabilities and safe unsupported actions.

Prompt Safety

Ask for examples, not authority escalation.

Prompts should not ask agents to approve claims, approve payouts, execute treasury movement, sign transactions inside SIMA, expose private keys, or treat quote/evidence/certificate records as guaranteed compensation.