Agent Skills
› opensquilla/opensquilla
› meta-scheduled-morning-digest
meta-scheduled-morning-digest
GitHub聚合天气与新闻，生成结构化晨间摘要并保存至长期记忆。适用于个人资讯定制与信息归档场景。
触发场景
用户请求每日早报
定时任务触发晨间信息汇总
安装
npx skills add opensquilla/opensquilla --skill meta-scheduled-morning-digest -g -y
SKILL.md
Frontmatter
{
"kind": "meta",
"name": "meta-scheduled-morning-digest",
"always": false,
"triggers": [
"晨报",
"morning digest",
"每日简报"
],
"provenance": {
"origin": "opensquilla-original",
"license": "Apache-2.0"
},
"composition": {
"steps": [
{
"id": "weather",
"with": {
"location": "{{ inputs.user_message | xml_escape | truncate(128) }}"
},
"skill": "weather"
},
{
"id": "news",
"with": {
"query": "{{ inputs.user_message | xml_escape | truncate(256) }} latest news",
"engines": [
"brave",
"duckduckgo"
],
"max_results": 5
},
"skill": "multi-search-engine"
},
{
"id": "digest",
"with": {
"text": "Weather:\n{{ outputs.weather }}\n\nNews:\n{{ outputs.news }}",
"style": "bulleted",
"max_words": 500
},
"skill": "summarize",
"depends_on": [
"weather",
"news"
]
},
{
"id": "memorize",
"kind": "tool_call",
"tool": "memory_save",
"tool_args": {
"mode": "append",
"path": "memory\/morning-digest.md",
"content": "{{ outputs.digest }}"
},
"depends_on": [
"digest"
],
"tool_allowlist": [
"memory_save"
]
}
]
},
"description": "Compose a morning digest combining local weather, news for the user's interest topic, a structured summary, and a memory note.",
"meta_priority": 40
}
Scheduled Morning Digest (Meta-Skill)
Pulls today's weather and news on a topic, summarizes them, and records
the digest in long-term memory for later recall. The MVP runs once per
invocation; recurring scheduling is left to the host (cron skill or
external scheduler) and is intentionally out of scope here.
Fallback
Have the LLM call weather, then multi-search-engine, summarize, and
finally memory_save manually.
版本历史
- 7f72a32 当前 2026-07-05 18:41
