The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill communicates with external Carbium endpoints for Solana RPC, WebSocket, and gRPC functionality, including rpc.carbium.io, api.carbium.io, wss-rpc.carbium.io, and grpc.carbium.io. These connections are necessary for the skill's primary function of blockchain interaction.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection because it ingests untrusted data from the Solana blockchain that could contain adversarial instructions designed to influence the agent. ● Ingestion points: Untrusted data enters the agent context through connection.onAccountChange and connection.onLogs in examples/websocket-monitor/README.md, as well as via the gRPC transactionNotification stream in examples/grpc-stream/README.md. ● Boundary markers: The provided examples do not implement boundary markers or instructions to ignore embedded commands within the ingested blockchain data. ● Capability inventory: The skill possesses the capability to perform network requests and to sign and submit blockchain transactions using the @solana/web3.js library. ● Sanitization: The skill does not demonstrate any sanitization, filtering, or validation of transaction logs or account data content before it is processed.

carbium