Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问许可证需确认审计通过

documenterdocumenter 搜索

Agent Skill

documenter 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

388

周安装

16

GitHub Stars

3

下载量

127
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:documenter(documenter 搜索)
来源仓库:https://github.com/n-n-code/n-n-code-skills
仓库路径:skills/documenter
安装命令:
npx skills add https://github.com/n-n-code/n-n-code-skills --skill documenter
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/n-n-code/n-n-code-skills --skill documenter

简介

用于查找、检索和筛选相关信息,支持关键词匹配与来源线索定位。

  • 适合在技术写作、API 注释和文档流程中快速获取候选内容或参考资料。
  • 优先识别受众类型和文档类型,避免混合多种模式,保持单一职责原则。
  • 通过 GitHub 安装,建议确认是否会触发联网或文件读写后再启用。
  • documenter 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

SKILL.md

Documenter

Documentation workflow for repo docs, technical writing, and inline API comments.

Priorities

  • keep docs accurate, scannable, and current
  • identify audience, document type, and source of truth before drafting
  • prefer examples and concrete structures over vague prose
  • match README/docs to actual build, test, install, and release behavior
  • keep public API comments complete enough for generated docs
  • keep one document one job; link to neighboring docs instead of mixing modes

Not For

  • binary document manipulation such as .docx, .pdf, .pptx, or .xlsx
  • tiny typo, grammar, or copy edits that do not require document-type decisions, repo-truth checks, or restructuring
  • regulatory document-control systems, compliance programs, or quality records
  • repo-specific doc-site pipelines unless the repo already contains them
  • intensive coauthoring sessions for large specs or proposals use documenter-coauthoring for that mode

Core Workflow

  1. Classify the request before writing:

- audience: user, contributor, maintainer, reviewer, or agent - document type: README, spec, ADR, tutorial, how-to, reference, explanation, API doc, code comment, changelog, release note - output target: existing file, new file, or inline content

  1. Read the current doc plus local source of truth in code, config, build scripts, tests, release files, and existing repo guidance.
  2. Pick the structure that fits the job:

- small edit: update only affected sections - large or new doc: outline first, then draft section by section

  1. Draft in compact Markdown with active voice, present tense, and direct wording. Explain before code when the reader needs context.
  2. Validate every claim you can:

- run or inspect commands and examples when feasible - check links, paths, identifiers, and filenames - if examples cannot be validated, say so instead of implying they were checked

  1. Tighten before finishing: remove duplication, filler, and mixed-purpose sections. Cross-link instead of repeating background.

Bundled Reference

Load references/doc-templates.md when the task needs a compact starting structure for README sections, ADRs, endpoint docs, public API comments, or changelog entries.

Document Type Rules

Use Diataxis when the request is general documentation rather than a repo convention with its own fixed format.

  • tutorial: for first-time learning by doing; every step should produce a visible result
  • how-to: for solving one known task; assume baseline knowledge and skip theory
  • reference: for facts, options, endpoints, flags, or API contracts; optimize for fast lookup
  • explanation: for trade-offs, concepts, rationale, and "why"

Keep each document a single dominant type. Link to related docs instead of mixing tutorial steps, reference tables, and conceptual background in one page.

README And Project Docs Rules

  • answer what it is, how to build it, how to test it, and how to release it
  • keep quick start runnable in minutes
  • keep top-level structure compact: title, one-line purpose, quick start, key commands, important constraints
  • avoid stale feature claims and aspirational behavior presented as shipped
  • use tables for config, support matrices, or command summaries when they make scanning faster

Specs, ADRs, And Change Docs Rules

  • specs and proposals should state audience, problem, constraints, decision, and open questions
  • ADRs should separate context, decision, and consequences
  • changelogs and release notes should emphasize user-visible changes, breaking changes, and upgrade action
  • if a decision depends on local politics or history, state the consequence plainly instead of hinting at it

Code Comment Rules

  • comment why, contracts, and non-obvious behavior
  • do not narrate obvious statements line by line
  • keep public API comments complete enough for Doxygen/docs generation
  • include params, return values, errors, and examples when the API surface needs them
  • do not invent API behavior that code does not implement

Documentation Hygiene Rules

  • put durable documentation where the repo already expects it; do not force a generic docs/ layout onto repos that use a different structure
  • make docs discoverable: link new durable docs from the nearest index, README, sidebar, or parent doc when the repo has one
  • prefer contextual links over vague text such as "click here"
  • use relative links for repo-local docs unless the repo clearly prefers another pattern
  • keep time-sensitive status, progress snapshots, and point-in-time test output out of durable docs unless the file is explicitly meant to record history, as in changelogs or ADRs
  • when writing docs ahead of implementation, mark planned interfaces clearly and replace placeholders with real behavior once the code exists

Style Rules

  • prefer simple, concrete words: use, add, run, because
  • prefer active voice; passive is acceptable when the actor is irrelevant
  • keep paragraphs short; split when a reader would need to re-parse the sentence
  • use direct address in tutorials and how-to guides; use neutral voice in references
  • avoid throat-clearing such as "this section describes" unless it adds meaning

Validation Rules

  • examples should be runnable, inspectable, or clearly marked illustrative
  • if public headers or external APIs change, update API comments immediately
  • keep checklists actionable
  • keep agent-facing docs aligned with repo guidance and local skills
  • for AI-facing docs, prefer clear heading hierarchy, self-contained sections, and direct file/path references

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.07%
按下载量换算47

Claude

28.57%
按下载量换算36

Cursor

18.39%
按下载量换算23

Gemini CLI

8.72%
按下载量换算11

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

external-service

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills