Security audit
Security checks for vulnerabilities and agentic risk
This skill is a narrow Binance content-drafting helper that uses disclosed external APIs and does not install code, persist itself, read local data, or take account actions.
Install only if you are comfortable sending the requested topic angle and any share-card fields to the disclosed Vercel-hosted service. Do not include non-public portfolio details, private trading notes, or credentials, and review any generated Binance Square draft before posting it yourself.
The description explicitly states the skill generates Binance Square Chinese drafts, and later guidance continues to assume Chinese-language output as the default behavior. This is a natural-language locale policy issue because the file imposes a specific language without user opt-in or a documented reason that the skill is limited to that locale.
The skill instructs callers to send requests and share-image parameters to an external Vercel-hosted service but does not disclose this data transfer or its privacy implications. Even if the current example payload is minimal, user prompts, market-analysis context, and generated share content may be transmitted to a third party without informed consent.
The skill includes a direct POST to an external endpoint, creating an external transmission path for user-influenced data. In this context, the danger is elevated because the skill is designed to generate analysis and publication materials; if later populated with user context or watchlist details, those inputs could be sent off-platform without adequate notice or minimization.
调用：
curl -sS -X POST https://binance-narrative-newsroom.vercel.app/api/newsroom \
-H 'Content-Type: application/json' \
-d '{"audience":"SQUARE","angle":"MORNING"}'
The skill enables implicit invocation without any visible activation constraints, scoping rules, or narrow trigger conditions. That means the agent may auto-select this skill based on vague relevance, causing unreviewed access to its prompt behavior and increasing the chance of prompt-routing abuse, accidental invocation, or misuse in contexts the user did not clearly intend.
No suspicious patterns detected.