collectorcrypt

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill includes examples that require embedding bearer/access/refresh tokens into HTTP headers or curl commands (e.g., "Authorization: Bearer " and SIWS token exchanges), which would force an LLM-driven agent to handle and potentially output secret values verbatim.

MEDIUM W009: Direct money access capability detected (payment gateways, crypto, banking).

• Direct money access detected (high risk: 1.00). The skill docs explicitly expose crypto financial operations: endpoints to buy/sell NFTs (USDC), build unsigned Solana transactions, generate gacha purchase transactions, submit/supply signed transactions (broadcast/submitTransaction), escrow/offer escrow management, buybacks, and an on-chain swap program. It describes signing flows, broadcasting transactions, and required wallet signatures — i.e., concrete APIs to move funds / execute on-chain transfers. This is direct financial execution capability.