Home / All Skills / Workflow & Core Tools / ljg-is
English (EN)简体中文 (ZH)日本語 (JA)
Workflow & Core Tools · lijigang/ljg-skills

ljg-is

Turns a noun or concept into an actionable understanding: first clarifying what it is in plain language and how it differs from adjacent concepts, then explaining how it operates and what judgments it reshapes, and finally providing practical leverage for acting on it. USE WHEN users ask what X is, how to understand it, what it implies, why it matters, or how to judge or use it in the future; especially suitable for technical concepts, systems, products, roles, methods, specifications, and practices. NOT FOR purely looking up facts, formula proofs, step-by-step how-to tutorials, deep longitudinal causal analysis (use ljg-think), or motif structures (use ljg-structure).

L4 High-Privilege · Graded auth & full auditV1 Commit-PinnedExecute CommandWrite FileDual-pass Audit
Author lijigangGitHub lijigang/ljg-skills ↗License MIT（仓库根 LICENSE）Commit 9e75497a95
Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

Core mechanism: Bridges "what it is" (noun) and "how it operates" (verb) into an actionable understanding.

skills/ljg-is/SKILL.md
`ljg-is` 把两者接起来：名词回答「它是什么」，让概念在认知地图里站稳；动词回答「它怎样起作用」，让概念在现实里动起来。
What is happening here: Uses the two-stage connection of "grounding the noun + activating the verb" as the main thread throughout the text.

The four guiding questions form a path of understanding, not four headings in the body text.

skills/ljg-is/SKILL.md
这四个问题是一条理解路径，不是四个正文栏目。
What is happening here: Opposes turning "what it is / how it operates / cognitive shift / action guidance" into a checklist-style structure.

Zhao Tingyang's "verb-first" thinking is a conditional lens, not a mandatory plotline.

skills/ljg-is/SKILL.md
赵汀阳的「动词」思想在这里仍然重要，但它是一副有条件的透镜。
What is happening here: Only systems, platforms, and roles query the selector and reflexivity; technical objects focus solely on their own internal operations.

Workflow Routing delegates execution to the UnderstandInUse workflow document.

skills/ljg-is/SKILL.md
把「是什么」与「怎样运作」接成可辨认、可判断、可行动的 Org 解读
What is happening here: The routing table maps the single workflow to Workflows/TraceCreation.md.

The finished standard requires four consecutive shifts in the reader: positioning, restating the mechanism, correcting prior judgments, and translating into concrete action.

skills/ljg-is/SKILL.md
能用一句普通话说出 X 属于什么，以及它与最容易混淆的对象差在哪里。定义要给边界，不能只给比喻或用途。
What is happening here: Breaks "understanding" down into four verifiable stages, requiring definitions to provide boundaries rather than merely use cases.

The default artifact is Org, written to ~/Context with a fixed schema ljg-is-v5.

skills/ljg-is/SKILL.md
新成品使用 `ljg-is-v5` schema。
What is happening here: Fixes the persistent storage path and backend metadata contract to facilitate indexing and validation.

The five backend items (definition, operation, cognitive correction, action judgment, basis and boundaries) are analytical preparations, not the body table of contents.

skills/ljg-is/Workflows/TraceCreation.md
这五件事是分析者的准备，不是正文目录。
What is happening here: Requires clarifying the five items in the background while prohibiting field-like headers and synonym-based itemized filling in the body text.

Acceptance contract: Run the local Bun validator after saving; even if validation passes, a full read-back is still required.

skills/ljg-is/Workflows/TraceCreation.md
bun ~/.agents/skills/ljg-is/Tools/ValidateNote.ts "<Org 文件路径>"
What is happening here: Separates automated checks (six fields, headers, paragraph count, readability) from manual human read-back into two distinct gates.

Judgment boundary: Technical objects do not need fictitious creators; describe only the mechanisms they actually possess.

skills/ljg-is/SKILL.md
**动词不等于社会创制。** 技术对象可以通过计算、映射、比较、压缩或递归起作用，不必虚构一个创造者故事。
What is happening here: Prevents forcing technical descriptions into institutional critique or philosophical twists.

2Core Capabilities

01Write nouns/concepts into an actionable understanding (what it is → how it works → cognitive correction → actionable leverage)
02Generate Org notes according to the v5 contract and write to ~/Context by default
03Run the local validator to mechanically check the six backend fields, content-driven headers, paragraph counts, and basic readability
04Intercept format anti-patterns: bullet points, field-like headers, overly dense enumeration marks, excessive semicolons, and run-on sentences
05Read legacy schemas (v2/v3/v4/legacy) with backward compatibility, allowing older notes to still be validated
06Ingest attached files, books, articles, URLs, and conversation materials as grounding, distinguishing facts from inferences
07Execute four-test read-back (noun, verb, cognition, action) and dedicated conclusion-scale checks
08Handle failure boundaries: narrow the scope, supply missing materials, or maintain simplicity when definitions, mechanisms, or cognitive shifts cannot be found

3External Dependencies

KindDependency / Endpoint
clibun（Bun 运行时，执行校验器与自检测试）

4Recommended Governance & Evidence

Suggested Controls for Safe Operation
For Agent Runtime
Graded/per-use authorization, sandbox recommended
For Administrators
Approval workflow, rollback support, full audit
Evidence & Triggers (File Paths & Excerpts)
  • scripts/: Execute local script or command: bun ~/.agents/skills/ljg-is/Tools/ValidateNote.ts "<Org file path>" (Acceptance contract: run after saving; output...
  • 本地工作目录: File I/O and artifact output: ~/Context/{timestamp}--理解-{target_fragment}__is.org (generated by copying Template.org; except when the user explicitly specifies "analysis only", "do not persist", or rea...
Audit Notes: Local script tier: This skill includes a Bun/TypeScript validator (Tools/ValidateNote.ts + tests + Template.org). The acceptance contract requires executing bun for validation and persisting org files to ~/Context, categorizing it as "local scripts/CLI and local I/O only", evaluated as blue. The script only imports node:path and performs plain text validation, with no network access, no credentials, and no subprocess escapes; URL detections in scanning are test fixtures pointing to example.com, and credential detections are false positives triggered by template string ${...} interpolations, neither representing real external calls, so no tier escalation is warranted.

5Second-Pass Verification

  • [ok] 两个执行点真实存在 — TraceCreation.md 验收合同给出 bun .../ValidateNote.ts 命令，ValidateNote.help.md 给出同命令与 bun test；脚本首行 #!/usr/bin/env bun。
  • [ok] 脚本无网络/凭证/写盘 — ValidateNote.ts 仅 import { basename } from "node:path"，无 fetch/http/fs 写/process.env/child_process；扫描两类命中经核为误报。
  • [discrepancy] 扫描「凭證/環境變數」14 条为误报 — 命中项均为 ValidateNote.test.ts / ValidateNote.ts 里的 JS 模板字符串 ${DEFINITION}、${V3_HEADINGS.join(...)} 等插值，不是环境变量或凭证读取；credential_reads 据实留空。
  • [discrepancy] 扫描「URL/端點」1 条为测试样例 — 唯一命中是 test.ts 用 [网约车](https://example.com) 验证「Org 中出现 Markdown 链接」会被拒，非网络请求；network_calls 留空。
  • [ok] 成品路径与 schema 一致 — SKILL.md 与 TraceCreation.md 一致写 ~/Context/{时间戳}--理解-{目标片段}__is.org，schema 固定 ljg-is-v5、标签 :is:act:。
  • [ok] v5 后台六字段一致 — definition/operation/recognition/guidance/basis/falsifier 在 SKILL.md、help.md、TraceCreation.md、Template.org 与测试中一致。
  • [ok] 判级归属 — 运行 bun 与本地写盘均为本 skill 自身脚本/本地动作，按范式判 blue；无凭证/网络故不升档。

6Verdict & Scenarios

  • Breaks "understanding" into a verifiable four-stage path (definition, mechanism, cognitive correction, action), accompanied by a deterministic validator to close the loop.
  • Codifies the rejection of generic advice into an enforceable check: fails validation if action guidance still holds after removing the mechanism section.
  • Separates backend fields from foreground prose, prohibiting field-like headers in the body to prevent the analytical framework from leaking into a checklist.
  • Supports backward compatibility across historical schemas (v2–v5 / legacy), ensuring older notes remain readable and migration does not break existing archives.
    • Best for: Suitable for writers and researchers who want to articulate technical concepts, institutions, products, roles, methods, or specifications "to the point of usability," especially org-mode users who prioritize definition boundaries, underlying mechanisms, and actionable judgments.
    Not for: Not suitable for: scenarios requiring only formal definitions, formula proofs, or term translations; users seeking step-by-step how-to tutorials; deep longitudinal causal analysis (use ljg-think) or motif structures (use ljg-structure); and host environments lacking bun that cannot run the local validator.
    Install Agent-Ready
    ① Verified Mirror 2026-10-10
    Option A · Direct DownloadDownload ljg-is.tar.gz
    sha256: 64b38a69ab8a2fac…
    Option B · Copy Guide for Agent
    install.json
    Agents automatically download and verify mirrors from this JSON guide.
    ② Upstream GitHub
    GitHub Repo ↗
    Pinned to commit 9e75497a95.