The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx @bnb-chain/mcp@latest to dynamically load and run the Model Context Protocol server. This package is managed by the official bnb-chain organization and is considered a trusted vendor resource.\n- [REMOTE_CODE_EXECUTION]: The skill invokes npx to download and execute code from the npm registry at runtime. Since the source is the official bnb-chain vendor, this is a standard and safe deployment method.\n- [COMMAND_EXECUTION]: The server is started using the npx command to initialize the MCP environment.\n- [CREDENTIALS_UNSAFE]: The documentation provides safe guidance for managing a PRIVATE_KEY via environment variables, correctly advising against hardcoding or exposure in logs.\n- [PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection by reading data from the blockchain.\n
Ingestion points: Tools such as get_transaction, read_contract, and gnfd_download_object ingest data from external sources.\n
Boundary markers: No explicit delimiters are used for processed data.\n
Capability inventory: The skill includes sensitive operations such as transfer_native_token, write_contract, and local file system access via Greenfield tools.\n
Sanitization: Not explicitly documented; however, the skill mandates user confirmation and network selection for all state-changing operations, significantly mitigating the risk.

bnbchain-mcp