The Agent Skills Directory

[SAFE]: The skill provides comprehensive documentation and functional templates for interacting with Curve Finance. The implementation prioritizes user safety by enforcing slippage tolerances and verifying contract state before execution.
[CREDENTIALS_UNSAFE]: No hardcoded secrets or sensitive credentials were found. The skill correctly utilizes environment variables (process.env.PRIVATE_KEY, process.env.RPC_URL) for secret management, which is a standard secure practice.
[DATA_EXFILTRATION]: There is no evidence of data exfiltration. All network traffic is directed toward the blockchain RPC endpoint provided by the user, and no sensitive local data is accessed or transmitted externally.
[PROMPT_INJECTION]: The skill instructions and metadata are focused on technical integration and do not contain any patterns intended to override agent safety guidelines or manipulate LLM behavior.
[EXTERNAL_DOWNLOADS]: The skill references the well-known and trusted viem library from the official NPM registry. No suspicious or unverified third-party dependencies are included.
[COMMAND_EXECUTION]: The code is restricted to TypeScript logic using standard libraries; it does not invoke shell commands, spawn subprocesses, or perform any high-risk system operations.

curve