The Agent Skills Directory

[SAFE]: The skill is designed purely for simulation. Analysis of the core scripts (flash_simulator.py, strategy_engine.py, profit_calculator.py) confirms that the logic focuses on mathematical modeling of arbitrage, liquidations, and fees using mock data or user-supplied parameters.
[SAFE]: The skill demonstrates safe credential management practices. It avoids hardcoding secrets and provides instructions for users to manage their own RPC endpoints and API keys through environment variables or local configuration files.
[SAFE]: External network references in config/settings.yaml and documentation target well-known, legitimate blockchain infrastructure providers (e.g., Ankr, LlamaRPC, Flashbots, Chainlist). These are standard and reputable resources within the DeFi development ecosystem.
[SAFE]: All identified Python dependencies (web3, httpx, pydantic, rich) are standard, widely-used, and reputable libraries for blockchain interaction and CLI application development.
[SAFE]: The skill implements restricted execution via scoped tool permissions (Bash(crypto:flashloan-*)) and maintains a read-only architectural posture, as described in the ARD.md and PRD.md documentation.

simulating-flash-loans