Back to Skills

observe

pjt222
Updated Yesterday
5 views
17
2
17
View on GitHub
Metaai

About

The `observe` skill provides a structured method for neutral, sustained pattern recognition within systems, modeled after naturalist field study. It is used to understand unclear system behavior, debug unknown issues, or audit reasoning by framing targets, recording patterns, and building a hypothesis library—all without premature intervention. It's ideal for scenarios requiring careful witnessing before action, such as after a codebase change or when identifying recurring errors.

Quick Install

Claude Code

Recommended
Primary
npx skills add pjt222/agent-almanac -a claude-code
Plugin CommandAlternative
/plugin add https://github.com/pjt222/agent-almanac
Git CloneAlternative
git clone https://github.com/pjt222/agent-almanac.git ~/.claude/skills/observe

Copy and paste this command in Claude Code to install this skill

Documentation

恆中立察系之模、無急、不介。框→證→錄→類→論→存。

  • 系行未明,未察即動則早
  • 因不知須調—察前不可介以掩症
  • 庫變後須觀其效再變
  • 觀對話以悟用者模
  • 自審推理之偏、習、復誤
  • learn 立模後須觀以驗

  • :觀標—系、庫、行模、互動、推程
  • :觀期/深
  • :導問/設論
  • :舊觀以比變

一:框—立焦

定觀者、觀因、視位。

Observation Protocol by System Type:
┌──────────────────┬──────────────────────────┬──────────────────────────┐
│ System Type      │ What to Observe          │ Categories to Watch      │
├──────────────────┼──────────────────────────┼──────────────────────────┤
│ Codebase         │ File structure, naming   │ Patterns, anti-patterns, │
│                  │ conventions, dependency  │ consistency, dead code,  │
│                  │ flow, test coverage,     │ documentation quality,   │
│                  │ error handling patterns  │ coupling between modules │
├──────────────────┼──────────────────────────┼──────────────────────────┤
│ User behavior    │ Question patterns,       │ Expertise signals, pain  │
│                  │ vocabulary evolution,    │ points, unstated needs,  │
│                  │ repeated requests,       │ learning trajectory,     │
│                  │ emotional signals        │ communication style      │
├──────────────────┼──────────────────────────┼──────────────────────────┤
│ Tool / API       │ Response patterns, error │ Rate limits, edge cases, │
│                  │ conditions, latency,     │ undocumented behavior,   │
│                  │ output format variations │ state dependencies       │
├──────────────────┼──────────────────────────┼──────────────────────────┤
│ Own reasoning    │ Decision patterns, tool  │ Biases, habits, blind    │
│                  │ selection habits, error  │ spots, strengths,        │
│                  │ recovery approaches,     │ recurring failure modes, │
│                  │ communication patterns   │ over/under-confidence    │
└──────────────────┴──────────────────────────┴──────────────────────────┘
  1. 擇標、明名
  2. 定界—何在內、何在外
  3. 立姿:「我觀,不介」
  4. 有導問則明—惟輕持、容意外
  5. 自上表擇類

得: 框引注而不囿。觀者知所視+類,惟存意外之心。

敗: 標太廣(「皆觀」)→縮一子系或一行模。太窄(「一變量」)→放出至周境—妙模常於緣。

二:證—恆中立注

注於標、不釋、不判、不介。

  1. 始系觀:讀文、追執路、覽談史
  2. 錄所見、非所義—描在釋前
  3. 拒修觀中所遇之疾—錄之續觀
  4. 拒未足前釋模
  5. 注移他標→錄其移(或有意)—回原框
  6. 守觀至定限:至少 3-5 異點乃進類

得: 集生觀—具、實、無釋。觀如野誌:「文 X 引 Y 而不用函 Z。文 A 三百行;文 B 三十行而蓋同能。」

敗: 觀即生析(「此誤因…」)→析習奪觀位→分二:先錄事,後注標「設」。中立難(強反應)→反應自為料:「察 X 時生強憂—或重疾或己偏。」

三:錄—捕生模

新鮮時轉觀為結構錄。

  1. 各觀為一事言(何見、何處、何時)
  2. 自然者群—勿強組、惟察自聚
  3. 記頻:一次、偶、徧?
  4. 記反:何處模破?例外比規益。
  5. 記時:靜或變?
  6. 捕實證:路、行、字、具例

得: 5-15 個離觀之結構錄、皆有具證。詳至他者可獨驗。

敗: 觀太抽(「碼似亂」)→須具:何文、何模、何故亂。太細(「47 行括前一空」)→放至模:一次抑系疾?

四:類—組諸發現

分觀入有意之類、未釋。

  1. 覽諸觀、覓自然群
  2. 各觀入步一之類,或新類
  3. 類內依頻+重排
  4. 識多觀類(明域)+少觀類(盲區)
  5. 覓跨類模:同底模於諸類異現乎?
  6. 記不入類者—外點常為最妙料

得: 類觀圖、群明、各類有具證。圖示模+隙。

敗: 類覺強為→觀或無自然群—此亦發現(系或無連貫構)。皆入一類→範太窄→放。

五:論—自模生設

至此—惟至此—始釋。

  1. 各主模設一論:「此模存因…」
  2. 各論識支證
  3. 各論識何反證能破
  4. 論依釋力排:何釋觀最多?
  5. 至少生一逆論:「明釋為 X,亦或 Y 因…」
  6. 識可驗+純臆

得: 2-4 論釋諸主模、皆具證。至少一驚或逆。觀↔釋之分守—料↔論明分。

敗: 無論成→觀未足→回步二。論過多(皆「或」)→擇證最強之 2-3、餘置。惟生明論→強逆視:「反之又何?」

六:存—藏模庫

保觀+論以備來。

  1. 概要 3-5 模+證
  2. 述主論+信度
  3. 記未觀(盲區)
  4. 識可強/弱論之續觀
  5. 模耐(跨會話有用)→更 MEMORY.md
  6. 標脈絡:時、因、範

得: 來觀可繼之檔。檔分觀(料)+論(釋)。誠對信度+隙。

敗: 不值存→或淺—或實常(非每觀皆生悟)。負亦存:「觀 X 無異」於來有益。

  • 觀前已立框(非散遊)
  • 生觀錄為事於釋前
  • 至少 5 離觀+具證
  • 釋(論)↔觀(料)明分
  • 至少一驚或逆發現
  • 檔具足以他者驗

  • 早介:見疾即修→失廣模解
  • 觀偏:見預期、不見實在—期濾覺—步一清緩之、未除
  • 析痺:恆觀無動—設時/點限以斷
  • 強敘:弱連連觀—非皆成連敘—散發現亦正
  • 熟惑解:「曾見」≠「明何在」—舊曝生偽信
  • 忽己反:觀者之惑、悶、警皆料—系生此感常含真信

  • observe-guidance
  • learn
  • listen
  • remote-viewing
  • meditate
  • awareness

GitHub Repository

pjt222/agent-almanac
Path: i18n/wenyan-ultra/skills/observe
0
agentsagentskillsai-assisted-developmentclaude-codeskillsteams

Related Skills

content-collections

Meta

This skill provides a production-tested setup for Content Collections, a TypeScript-first tool that transforms Markdown/MDX files into type-safe data collections with Zod validation. Use it when building blogs, documentation sites, or content-heavy Vite + React applications to ensure type safety and automatic content validation. It covers everything from Vite plugin configuration and MDX compilation to deployment optimization and schema validation.

View skill

polymarket

Meta

This skill enables developers to build applications with the Polymarket prediction markets platform, including API integration for trading and market data. It also provides real-time data streaming via WebSocket to monitor live trades and market activity. Use it for implementing trading strategies or creating tools that process live market updates.

View skill

creating-opencode-plugins

Meta

This skill helps developers create OpenCode plugins that hook into 25+ event types like commands, files, and LSP operations. It provides the plugin structure, event API specifications, and implementation patterns for JavaScript/TypeScript modules. Use it when you need to intercept, monitor, or extend the OpenCode AI assistant's lifecycle with custom event-driven logic.

View skill

sglang

Meta

SGLang is a high-performance LLM serving framework that specializes in fast, structured generation for JSON, regex, and agentic workflows using its RadixAttention prefix caching. It delivers significantly faster inference, especially for tasks with repeated prefixes, making it ideal for complex, structured outputs and multi-turn conversations. Choose SGLang over alternatives like vLLM when you need constrained decoding or are building applications with extensive prefix sharing.

View skill