pear-protocol

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt shows examples that embed secrets verbatim into CLI/json requests (e.g., passing '' to gateway.py and the explicit POST /trade-accounts payload with "signer_key" containing the private key), which would force an agent to include secret values in its outputs/commands.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (low risk: 0.10). Runtime path ingests only provider-served market/account data over HTTP (e.g., GET /markets via scripts/gateway.py http_get and MCP tool results via scripts/pear_mcp.py), with no required workflow that reads outsider-authored free text without first selecting trusted items.

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

• Direct money access detected (high risk: 1.00). The skill explicitly exposes trading APIs and scripts that can place live market orders and manage trading accounts. It documents POST/write endpoints, a gateway script that performs authenticated writes, the ability to create read_write API keys that enable trading tools (plan/execute, execute_* = live), and the trade-account setup (including signing and POST /trade-accounts). These are specific financial execution capabilities (placing trades on Hyperliquid via Pear/MCP), not generic utilities.