菜单
Periodic review for loopany — daily, weekly, or monthly. Triggers: 'what's due today', 'what am I forgetting', 'what's slipping', 'weekly check', 'is the workspace healthy', 'is this still the right mission?', 'anything structural?', session start, daily/weekly/monthly cadence.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/superdesigndev/loopany --skill loopany-review命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Artifact lifecycle for loopany — the agent brain. Use when creating, modifying, or querying any artifact (task, signal, learning, skill-proposal, mission, person, note, brief, journal). Read this SKILL.md first for routing, then the relevant kind's Playbook. Triggers: any artifact CRUD, 'create a task', 'write a signal', 'dismiss signal', 'flip to done', 'what kind should this be', choosing a relation verb.
Self-improvement loop for loopany. Reads recent task outcomes + dismissed signals, discovers patterns, writes learning and skill-proposal artifacts. Also handles accepting/rejecting proposals. Triggers: 'reflect', 'what have we learned', 'improve yourself', ≥3 tasks done recently, 'accept <proposal-id>', 'reject <proposal-id>', 'review proposals', writing a learning or skill-proposal from any source.
Read FIRST in any loopany workspace. Routes user intent to the correct skill, enforces bootstrap, and orchestrates cross-skill chaining. Triggers: any loopany interaction, session start, ambiguous request, 'what should I do', multi-skill workflow.
基于 SOC 职业分类
正在显示 SKILL.md
| name | loopany-review |
| description | Periodic review for loopany — daily, weekly, or monthly. Triggers: 'what's due today', 'what am I forgetting', 'what's slipping', 'weekly check', 'is the workspace healthy', 'is this still the right mission?', 'anything structural?', session start, daily/weekly/monthly cadence. |
One skill, three frequencies. Each adds scope; all share the same closure gate.
| Freq | Scope | Trigger | Read |
|---|---|---|---|
| daily | checkAt = today | "what's due today", session start | references/daily.md |
| weekly | overdue + parking lots + doctor | "what's slipping", "weekly check" | references/weekly.md |
| monthly | mission alignment + structural drift | "right mission?", "anything structural?" | references/monthly.md |
Don't mix scopes. One week of overdue is a pattern; one day isn't.
Frequency-specific — see the reference file.
Read each: loopany artifact get <id>
checkAt with a one-line reason.checkAt forward with a reason.For learnings: loopany trace <learning-slug> --direction backward
Only bucket B. One line per item:
3 things need you today:
1. [<task-slug>] 2-week recheck — did $/session settle?
2. [<learning-slug>] "short deals close 2.5x faster" — still true?
3. [<signal-slug>] recurring churn signal, 3rd time. Upgrade?
Empty → say so and stop.
Route to kind playbooks in loopany-core:
../loopany-core/kinds/task.md § Playbook../loopany-core/kinds/learning.md § Playbook../loopany-core/kinds/signal.md § Playbook../loopany-core/kinds/mission.md § PlaybookEvery surfaced item must end the session in one of:
checkAt pushed with reason. "Not yet" isn't a reason.checkAt removed with note.No zombie items. A digest without state changes is noise.
DAILY: followups --due today → classify → surface → dispatch → close
WEEKLY: doctor + overdue + parking lots → classify → surface → close → feed reflect
MONTHLY: mission alignment + structural drift → surface → user decides
GATE: every surfaced item → resolved / deferred / retired