xiaoqi-token-stats

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

• Third-party content exposure detected (medium risk: 0.30). 技能在运行时会从本地路径 ~/.claude/projects/ 递归读取用户的 Claude Code 会话 JSONL（只要其中含有 assistant 消息的 model/usage/timestamp 字段），并把这些内容聚合后嵌入生成的 HTML（find_jsonl_files()→scan_all()→render_calendar_html()/render_html()）。

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

• Potentially malicious external URL detected (high risk: 0.70). The generated HTML in token_stats.py includes a script tag that will fetch and execute remote JavaScript from https://cdn.jsdelivr.net/npm/chart.js@4.4.7/dist/chart.umd.min.js when the report is opened (the skill auto-opens the HTML), so remote code is executed at runtime and the skill depends on it for chart rendering.