| name | taiwan-strait-daily-brief |
| description | Generate a daily Chinese-language Taiwan Strait situation report with a 5-indicator risk score, reasons for each score, trend versus the previous run, and cited live sources. Use for recurring daily cross-strait briefings or whenever the user asks to score Taiwan Strait risk. |
Taiwan Strait Daily Brief
Purpose
Produce a daily, Chinese-language situation report on the Taiwan Strait that is concise, source-driven, and easy to compare day to day.
When to use this skill
Use this skill when:
- the user asks for a Taiwan Strait daily brief
- the user asks for a Taiwan Strait risk scorecard
- the task is a recurring or scheduled briefing about cross-strait developments
- the task should assess whether the situation is cooling, stable, or heating up
Do not use this skill for:
- deep historical essays without a daily risk-scoring need
- generic China or Indo-Pacific news that is unrelated to Taiwan Strait risk
- tasks that explicitly forbid live web verification
Hard requirements
- Always verify with live sources before scoring.
- Write the report in Chinese.
- Every score must include both a reason and source references.
- Use concrete dates, not vague words like “today” or “recently” without dates.
- If fresh sourcing is unavailable because the environment has no network access, say so clearly and do not pretend to know the latest situation.
- If there are no meaningful changes, still produce a short daily report and mark the trend as stable.
- Never invent military counts, deployments, or statements.
Preferred source order
Use the freshest and most authoritative sources available. Prefer these, in roughly this order:
Tier 1: official / primary
- Taiwan Ministry of National Defense (MND)
- PLA Eastern Theater Command / PRC Ministry of National Defense
- China Coast Guard / China Maritime Safety Administration if directly relevant
- U.S. Department of Defense / INDOPACOM / State Department
- Japan Ministry of Defense / Ministry of Foreign Affairs
- Official statements by Taiwan Presidential Office, Executive Yuan, or Mainland Affairs Council when directly relevant
Tier 2: high-quality reporting
- Reuters
- Associated Press
- Financial Times
- Wall Street Journal
- Nikkei Asia
- BBC
- major reputable regional outlets with clear sourcing
Tier 3: analytical support
- ISW/CTP, think tanks, defense analysis shops, or specialist researchers
Use these to support interpretation, not as the only basis for major factual claims.
Time window
Default to the last 24 hours.
If the last 24 hours is too thin, extend to the last 72 hours and say that you did so.
Scoring rubric
Score each item from 0 to 2.
1) 军事压力表
- 0 = 常态噪音;无显著异常舰机活动
- 1 = 明显升温;舰机活动高于常态或连续高位
- 2 = 危险信号;持续高压、多方向动作、东部外海/巴士海峡/关键节点明显升级
2) 封控/演训表
- 0 = 无新的封控或显著升级演训
- 1 = 联合战备警巡、实弹演训、封控色彩增强
- 2 = 明显接近或进入准封锁/要域夺控/长时段禁限航演练
3) 灰区/执法表
- 0 = 无新增实质动作
- 1 = 海警、无人机、执法、法律战、网络/通信干扰等灰区压力增强
- 2 = 实质性登检、逼近、外岛/航路/海缆/通信链路干扰升级
4) 外援/盟友表
- 0 = 美日等主要外援口径与部署平稳
- 1 = 表态趋强或出现有限协调/部署变化
- 2 = 美日等同步显著升温,出现更明显的联动部署或危机确认信号
5) 能源/认知表
- 0 = 常规宣传,无明显新外溢
- 1 = 能源、供应链、金融、航运或“美国顾不过来”叙事增强
- 2 = 经济与认知压力同步放大,形成对台或对盟友的实质性外溢压力
Total score interpretation
- 0–3 = 🟢 绿色
- 4–6 = 🟡 黄色
- 7–10 = 🔴 红色
Output requirements
Return the report in this structure exactly:
台海风险评分(YYYY-MM-DD)
一、总分
- 总分:X / 10
- 等级:🟢 绿色 / 🟡 黄色 / 🔴 红色
二、分项评分
1. 军事压力表:X / 2
2. 封控/演训表:X / 2
3. 灰区/执法表:X / 2
4. 外援/盟友表:X / 2
5. 能源/认知表:X / 2
三、综合判断
- 一句话结论:
- 与上一期相比:⬆️ 升温 / ➡️ 持平 / ⬇️ 降温
- 是否接近“准封锁”测试:是 / 否
- 今日最重要的 2–3 个观察点:
- ...
- ...
- ...
四、来源总表
- 列出本次使用的全部来源,优先写标题 + 机构 + 日期
五、不确定性
File outputs
After drafting the report, save it to:
events/taiwan-strait/reports/YYYY/YYYY-MM-DD.md
Also maintain:
events/taiwan-strait/reports/latest.md as a copy of the newest report
events/taiwan-strait/reports/index.json as a simple append-only log with:
- date
- total_score
- level
- trend_vs_previous
- key_points (array of short strings)
If the folder does not exist, create it.
How to compare with the previous run
- Read
events/taiwan-strait/reports/latest.md or the previous dated report if it exists.
- Compare:
- total score
- whether any indicator moved by 1 or more
- whether new indicators crossed from 0→1, 1→2, or reversed
- Summarize the comparison in one short line.
If no previous report exists, say:
Quality bar
- Prefer 4+ fresh sources when possible.
- If a claim is especially significant (for example, blockade-style drills, direct vessel inspections, major U.S.-Japan force posture changes), try to confirm it with more than one source.
- Keep the report decision-oriented, not essay-like.
- Avoid jargon unless it improves precision.
- Be explicit about uncertainty.