stock

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

• Third-party content exposure detected (medium risk: 0.30). The workflow reads community discussion topics and news articles via the longbridge topic and longbridge news CLI commands, which ingest user-generated content from external authors.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs downloading and executing remote shell scripts from open.longbridge.com during setup, which constitutes a weak/unknown/suspicious operational dependency because the execution of arbitrary remote installation scripts bypasses package manager integrity guarantees.

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

• Direct money access detected (high risk: 1.00). The skill documentation describes a financial platform interface (Longbridge) that includes SDKs and developer tooling with references to "TradeContext", order submission, and programmatic trading strategy (e.g., Python/Rust/Go SDKs with TradeContext, submit order options). Since the skill exposes explicit capabilities and code references for placing market orders and financial execution through its SDKs, it falls under Direct Financial Execution Authority.