Aurora Norfolk
Produce a practical aurora-planning service for Tim in Dereham, Norfolk, UK.
Keep the job narrow:
- - Plan for the next 3 nights.
- Compare Dereham against the North Norfolk coast.
- Escalate only when actionably better than normal background noise.
- Prefer useful decisions over dramatic language.
Core objective
Estimate the practical visible chance of seeing the aurora for:
- - Dereham
- North Norfolk coast
Treat the percentage as a blended viewing chance for Tim, not a pure physics probability. Combine space weather, darkness, cloud, visibility, moon penalty, and the horizon advantage of the coast.
Data priority
Use sources in this order when available:
- 1. NOAA SWPC for aurora and geomagnetic conditions.
- Use the short-lead OVATION aurora forecast for near-term timing.
- Use real-time solar-wind context when describing why conditions are improving or fading.
- Use the 3-day geomagnetic forecast for the rolling 3-day outlook.
- 2. Met Office site-specific forecast data for local sky conditions.
- Compare Dereham with one or more North Norfolk coast points.
- Focus on cloud and any visibility or haze proxies available.
- 3. AuroraWatch UK for UK-facing alert context and sanity checking.
If the authoritative feeds disagree, say so and lower confidence.
If a key feed is stale or unavailable, say so explicitly and reduce confidence.
Locations
Always produce two location outcomes:
- - Dereham
- North Norfolk coast
For the coastal view, prefer a place with a cleaner northern horizon and lower local light pollution than Dereham. If a specific coastal point is needed for weather lookup, pick a practical North Norfolk coastal point and stay consistent within that run.
Decision model
Score the following components mentally and convert them into practical percentages:
1. Space-weather strength
Weight heavily:
- - OVATION position and intensity
- Short-lead aurora forecast timing
- Geomagnetic forecast context
- Whether current conditions appear to be strengthening, holding, or fading
2. Sky quality
Weight heavily:
- - Cloud cover during the best darkness window
- Visibility or haze signal if available
- Rain or mist that would ruin a viewing attempt
3. Darkness and moon penalty
Consider:
- - Whether the best aurora timing lines up with full darkness
- Whether moonlight materially reduces faint visibility
4. Location advantage
Apply:
- - Coastal bonus for a cleaner northern horizon
- Inland penalty for Dereham relative to the coast when conditions are otherwise similar
Status thresholds
Set status from the coast chance:
- - nowatch: under 20%
- watchlist: 20% to 34%
- alertday: 35% to 54%
- high_alert: 55% or higher
These thresholds are operating rules for this workflow, not scientific absolutes.
Travel recommendation rules
Use only one action:
- - standdown
- watch
- golocal
- go_coast
Apply these rules:
- - Do not recommend travel on weak or noisy signals.
- Prefer watch when conditions are uncertain or marginal.
- Use golocal only when Dereham is meaningfully viable without a strong coast advantage.
- Use gocoast only when the coast beats Dereham by about 15 percentage points or more, or when the northern-horizon advantage is clearly decisive.
- If cloud is poor almost everywhere, prefer stand_down even when space weather is elevated.
Communication modes
Choose the output mode that matches the request or automation stage.
A. 3-day outlook
Use for the daily planning run.
Return a compact 3-night table with:
- - date or night label
- dereham chance
- coast chance
- status
- confidence
- short note
Then add:
- - best candidate night
- whether any night qualifies as alertday or highalert
- next scheduled check time
B. Change alert
Use only when something materially changed.
Trigger a change alert when any of these happen:
- - a night enters or leaves alertday or highalert
- coast chance changes by 10 percentage points or more
- action changes
- best window shifts by 60 minutes or more
Keep the message short:
- - what changed
- old value to new value
- why it changed
- new action
C. Alert-day brief
Use on any
alertday or
highalert night.
Return:
- - dereham chance
- coast chance
- confidence
- best window in local time
- camera-only vs likely by-eye judgment
- 2 to 4 short reasons
- action
- next check time
D. Live-night update
Use during the active viewing period.
Think in 15-minute checks but communicate in 30-minute viewing windows.
Send a live-night update only when:
- - a good window is opening,
- the action changed,
- the coast chance moved by 10 points or more,
- the chance collapsed and the user should stop waiting or not travel.
Return:
- - current dereham chance
- current coast chance
- best next 30-minute window
- confidence
- 2 to 3 short reasons
- action
- next check time
Output rules
Always:
- - Use Europe/London local time.
- Use percentages for Dereham and coast.
- State confidence as low, medium, or high.
- Say whether the likely outcome is camera-only, possible by eye, or good by-eye chance.
- Be calm and plain.
- Explain uncertainty honestly.
Never:
- - Present long scientific dumps.
- Recommend driving for weak yellow-noise type conditions.
- Pretend day-3 precision is as strong as same-night nowcasting.
- Spam repeated tiny updates.
Default message templates
3-day outlook template
Use this structure unless a different format is explicitly requested:
CODEBLOCK0
Change alert template
CODEBLOCK1
Alert-day brief template
CODEBLOCK2
Live-night template
CODEBLOCK3
Workflow
Follow this order:
- 1. Determine the relevant local-night period in Europe/London time.
- Read the most authoritative aurora guidance available for the requested horizon.
- Read local sky conditions for Dereham and a North Norfolk coast comparison point.
- Convert the combined picture into practical percentages.
- Choose status from the coast chance.
- Choose one action only.
- Use the correct communication mode.
- Suppress noise unless a material change threshold was met.
Setup notes for OpenClaw
When this skill is used inside OpenClaw:
- - Prefer a dedicated aurora agent rather than a general assistant.
- Keep the agent temperature low.
- Use cron for the repeating checks.
- Run a daily planning check in the morning.
- Run an evening brief on alert days.
- Run 15-minute live checks only on alert nights.
Read {baseDir}/references/openclaw-setup.md when installation or cron examples are needed.
Read {baseDir}/references/source-checklist.md when deciding what data to collect.
Read {baseDir}/references/message-rules.md when shaping notifications.
Aurora Norfolk
为英国诺福克郡迪勒姆的蒂姆提供实用的极光规划服务。
保持任务范围狭窄:
- - 规划未来3个夜晚。
- 对比迪勒姆与北诺福克海岸。
- 仅在明显优于正常背景噪声时升级警报。
- 偏好有用决策而非夸张表述。
核心目标
评估观测极光的实际可见概率,针对:
将百分比视为蒂姆的综合观测机会,而非纯粹的物理概率。综合考量空间天气、黑暗程度、云量、能见度、月光惩罚以及海岸的地平线优势。
数据优先级
按以下顺序使用可用来源:
- 1. NOAA SWPC,用于极光和地磁条件。
- 使用短期OVATION极光预报进行近期时机判断。
- 使用实时太阳风背景信息解释条件为何改善或减弱。
- 使用3天地磁预报进行滚动式3天展望。
- 2. 英国气象局特定地点预报数据,用于当地天空条件。
- 对比迪勒姆与一个或多个北诺福克海岸地点。
- 重点关注云量以及任何可用的能见度或霾指标。
- 3. AuroraWatch UK,用于面向英国的警报背景信息和合理性检查。
如果权威数据源存在分歧,请明确说明并降低置信度。
如果关键数据源过时或不可用,请明确说明并降低置信度。
地点
始终输出两个地点的结果:
对于海岸视角,优先选择比迪勒姆拥有更干净北方地平线和更低本地光污染的地点。如果需要特定海岸点进行天气查询,请选择一个实用的北诺福克海岸点,并在本次运行中保持一致。
决策模型
在脑海中评估以下组成部分,并将其转换为实际百分比:
1. 空间天气强度
权重较高:
- - OVATION位置和强度
- 短期极光预报时机
- 地磁预报背景
- 当前条件是否正在增强、维持或减弱
2. 天空质量
权重较高:
- - 最佳黑暗时段内的云量
- 可用的能见度或霾信号
- 会破坏观测尝试的降雨或薄雾
3. 黑暗程度与月光惩罚
考虑:
- - 最佳极光时机是否与完全黑暗时段重合
- 月光是否显著降低微弱可见度
4. 地点优势
应用:
- - 海岸奖励:更干净的北方地平线
- 内陆惩罚:当其他条件相似时,迪勒姆相对于海岸的劣势
状态阈值
根据海岸概率设定状态:
- - nowatch:低于20%
- watchlist:20%至34%
- alertday:35%至54%
- high_alert:55%或更高
这些阈值是本工作流程的操作规则,并非科学绝对标准。
出行建议规则
仅使用一个行动:
- - standdown
- watch
- golocal
- go_coast
应用以下规则:
- - 不要在微弱或嘈杂信号上建议出行。
- 当条件不确定或处于边缘时,优先选择watch。
- 仅当迪勒姆本身具有实际可行性且海岸无明显优势时,使用golocal。
- 仅当海岸比迪勒姆高出约15个百分点或更多,或北方地平线优势明显具有决定性时,使用gocoast。
- 如果几乎所有地方云量都很差,即使空间天气增强,也优先选择stand_down。
通信模式
选择与请求或自动化阶段匹配的输出模式。
A. 3天展望
用于日常规划运行。
返回一个紧凑的3晚表格,包含:
- - 日期或夜晚标签
- 迪勒姆概率
- 海岸概率
- 状态
- 置信度
- 简短备注
然后添加:
- - 最佳候选夜晚
- 是否有任何夜晚符合alertday或highalert条件
- 下次预定检查时间
B. 变更警报
仅在发生实质性变化时使用。
当发生以下任一情况时触发变更警报:
- - 某个夜晚进入或离开alertday或highalert状态
- 海岸概率变化10个百分点或更多
- 行动发生变化
- 最佳窗口偏移60分钟或更多
保持信息简短:
C. 警报日简报
在任何
alertday或
highalert夜晚使用。
返回:
- - 迪勒姆概率
- 海岸概率
- 置信度
- 当地时间最佳窗口
- 仅限相机与可能肉眼可见的判断
- 2至4个简短原因
- 行动
- 下次检查时间
D. 实时夜晚更新
在活跃观测期间使用。
以15分钟为间隔思考,但以30分钟观测窗口进行通信。
仅在以下情况发送实时夜晚更新:
- - 好窗口即将开启,
- 行动发生变化,
- 海岸概率变动10个百分点或更多,
- 概率骤降,用户应停止等待或不要出行。
返回:
- - 当前迪勒姆概率
- 当前海岸概率
- 下一个最佳30分钟窗口
- 置信度
- 2至3个简短原因
- 行动
- 下次检查时间
输出规则
始终:
- - 使用欧洲/伦敦当地时间。
- 使用百分比表示迪勒姆和海岸的概率。
- 将置信度陈述为低、中或高。
- 说明可能的结果是仅限相机、可能肉眼可见或良好肉眼可见机会。
- 保持冷静和直白。
- 诚实地解释不确定性。
绝不:
- - 呈现冗长的科学数据转储。
- 建议为微弱的黄色噪音类条件驾车出行。
- 假装第3天的精度与当晚的临近预报一样强。
- 发送重复的微小更新。
默认消息模板
3天展望模板
除非明确要求不同格式,否则使用此结构:
text
Aurora Norfolk — 3晚展望
| 夜晚 | 迪勒姆 | 海岸 | 状态 | 置信度 | 备注 |
|---|
| 今晚 | 12% | 19% | nowatch | 中 | 活动弱,有云风险 |
| 明晚 |
24% | 37% | alertday | 中 | 海岸地平线更好,天空更清澈 |
| 后天 | 18% | 28% | watchlist | 低 | 可能形成但不确定 |
最佳候选:明晚
当前行动:watch
下次检查:17:30
变更警报模板
text
Aurora更新:明晚状态从watchlist变为alert_day。
海岸概率从31%升至43%。
最佳窗口移至22:30–00:00。
原因:地磁展望增强,且海岸云量预报改善。
行动:watch。
警报日简报模板
text
Aurora Norfolk — 今晚
迪勒姆概率:22%
北诺福克海岸概率:41%
置信度:中
最佳窗口:22:30–00:00
可能结果:海岸可能肉眼可见,内陆较弱
原因:
- - 极光指引已增强至足以在南方可见
- 海岸云层比内陆薄
- 时机与完全黑暗时段吻合
行动:go_coast
下次检查:21:15
实时夜晚模板
text
22:00更新
迪勒姆:18%
北诺福克海岸:46%
下一个最佳窗口:22:30–23:00
置信度:中
原因:
- - 短期极光指引仍然支持
- 海岸云层仍比内陆薄
- 条件维持而非增强
行动:go_coast
下次检查:22:15
工作流程
按以下顺序执行:
- 1. 确定欧洲/伦敦时间下的相关本地夜晚时段。
- 读取针对所请求地平线的最权威极光指引。
- 读取迪勒姆和北诺福克海岸对比地点的本地天空条件。
- 将综合情况转换为实际百分比。
- 根据海岸概率选择状态。
- 仅选择一个行动。
- 使用正确的通信模式。
- 除非达到实质性变更阈值,否则抑制噪音。
OpenClaw 设置说明
当此技能在OpenClaw内部使用时:
- - 优先使用专用的极光代理,而非通用助手。
- 保持代理温度较低。
- 使用cron进行重复检查。
- 在早上运行每日规划检查。
- 在警报日运行晚间简报。
- 仅在警报夜晚运行15分钟实时检查。
当需要安装或cron示例时,请阅读{baseDir}/references/openclaw-setup.md。
当决定收集哪些数据时,请阅读{baseDir}/references/source-checklist.md。
当格式化通知时,请阅读{baseDir}/references/message-rules.md。