weex-monitor-skill

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

• Third-party content exposure detected (high risk: 0.75). Runtime path confirm-text-live → _collect_live_account_payload() delegates to weex_trade_data_aggregator.py to fetch a live account risk/position snapshot, which is outsider-authored free-form JSON/prose from external vendor tooling and is then rendered into the agent’s LLM context via confirmation_text/thread_report.

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

• Direct money access detected (high risk: 1.00). This skill is specifically designed to monitor positions and trigger market-close orders. Its Task DSL includes an "action.type":"market_close" and the skill provides flows to "confirm-and-run-loop" and "run-live-once" which, when authorized, perform market-close execution (via preview-order and confirm-order). Although it delegates REST signing/credentials and the actual submission to weex-trader-skill, this skill explicitly orchestrates and can initiate live market order execution given user confirmation and matching execution authorization. That makes it a financial-execution-capable skill (market orders).