stock-announcement-analysis

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt includes an explicit Tushare Pro API token literal in example code (pro_api('c8dbb38...')) and shows embedding it directly in code/requests, which requires handling/outputting a secret verbatim and thus creates an exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 0.90). The document instructs deliberate scraping-evasion techniques (proxy pools, UA rotation) and includes a hard-coded Tushare API token plus guidance to redirect API calls to a non-official host, enabling credential exposure and potential unauthorized data exfiltration.

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

• Third-party content exposure detected (high risk: 0.85). 该工作流会通过 Hermes/browser 工具打开 CNINFO 公告详情页并下载/提取“公告PDF正文”的可读文本；这些公告正文属于第三方公开网页内容（非操作用户自写），从而在运行时进入代理的 LLM 上下文，构成外部自由文本注入风险。

HIGH W008: Secret detected in skill content (API keys, tokens, passwords).

• Secret detected (high risk: 1.00). 发现一个高熵、字面出现的可用API密钥：在 Tushare 示例代码中有 pro_api('c8dbb3833192a3e47991b1975ad02d95a6567988826e519ba76b0ef5')。该字符串为长、随机样式的字面值，符合“实际可用凭据”的定义，可能可用于访问 Tushare 服务，因此应视为秘密并标记。

其余内容均为可忽略项：

• URL、接口说明、环境变量名（如 AGENT_BROWSER_EXECUTABLE_PATH）不是密钥。
• 其他示例/说明性的简单字符串（如提示需配置 token，但未提供）、以及说明性账号/占位符均为文档占位或低熵示例，故忽略。