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

communicatecommunicate 搜索

Agent Skill

communicate 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

196

周安装

8

GitHub Stars

1

下载量

63
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:communicate(communicate 搜索)
来源仓库:https://github.com/synapseradio/ai-skills
仓库路径:skills/communicate
安装命令:
npx skills add https://github.com/synapseradio/ai-skills --skill communicate
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/synapseradio/ai-skills --skill communicate

简介

提供结构化写作技巧集合,涵盖清晰表达、诚实沟通与 engaging 叙述五大类方法。

  • 适用于撰写报告、提案或用户文档时提升逻辑性与可读性。
  • 支持按任务类型路由至特定工作流,也允许灵活调用单条技术 ad-hoc 使用。
  • 默认在模糊意图下采用通用技巧组合,复杂场景建议明确需求后切换专用模式。
  • 所有输出建议均基于预设分类体系,实际效果需结合目标读者反馈持续迭代。

SKILL.md

Communicate

Comprehensive written expression skill. 16 techniques across 5 categories (plus a tradition meta-lens and omit gate) for crafting prose that communicates clearly, honestly, engagingly, and appropriately for its audience and rhetorical tradition. Routes to structured workflows for common tasks or applies techniques ad-hoc.

Intent Routing

Detect user intent and route to the appropriate mode. When ambiguous, default to ad-hoc technique application. If scope becomes clear mid-conversation, route to the appropriate workflow.

Intent SignalRouteLoad
Improve/refine specific issues, apply techniques, fix passive voice, remove weasel wordsAd-hoc (use diagnostic tables below)Technique files as needed
"proofread", "check my writing", "clean this up", "edit this", "fix my draft", structured clarity auditClarify@./references/workflow-clarify.md
"make this as good as it can be", "full refinement", "polish thoroughly", "this needs to persuade"Redraft@./references/workflow-redraft.md
"help me write", "draft this", "compose", "I have something to say", blank-page creationCompose@./references/workflow-compose.md
"adapt for audience", "translate the structure", "why does my writing feel foreign?", cross-traditionStructure@./references/workflow-structure.md
"help me say this", "I can't find the right words", "in [language] we say...", cross-lingual bridgingBridge@./references/workflow-bridge.md

Bridge as interrupt: The Bridge workflow can interrupt any active workflow. When a thought resists English expression mid-workflow, load @./references/workflow-bridge.md for the stuck passage, then return to the active workflow.

Context Loading

Load only what the current route requires. Never load all references upfront.

RouteLoadDo NOT Load
Ad-hoctradition.md first, then technique files by diagnostic questionworkflow-*.md files
Clarifyworkflow-clarify.md (loads its own techniques by phase)Other workflow-*.md, index.md
Redraftworkflow-redraft.md (loads its own techniques by pass)Other workflow-*.md, index.md
Composeworkflow-compose.md (loads its own techniques by phase)Other workflow-*.md
Structureworkflow-structure.md, tradition-signatures.mdOther workflow-*.md
Bridgeworkflow-bridge.md, translanguaging-guide.md, language-patterns.mdOther workflow-*.md
Technique navigationindex.mdworkflow-*.md files

Tradition Meta-Lens and Omit Gate

Tradition. Load before any technique. Identify the rhetorical tradition operating in the prose before applying diagnostics — this prevents Anglo-American conventions from functioning as an invisible default and modifies how all downstream questions fire.

@./references/tradition.md

Why it matters: every diagnostic question below assumes a rhetorical context. "Can a reader follow linearly?" fails for Arabic parallel-restatement. "Does language commit?" fails for Japanese epistemic calibration. Tradition identification prevents misdiagnosis.

Omit gate. Before repairing any passage, ask: *Should this content exist at all?* Not every gap is a flaw — some silences are structural. Intentional omission sharpens focus by refusing to dilute a point. Strategic silence lets the reader complete the thought, which lands harder than spelling it out. Productive ambiguity preserves openness where premature precision would close inquiry. If the answer is "this content shouldn't exist," the repair is deletion, not technique application.

Orientation

Treat reader attention as finite. Interrogate each word: does it earn its place?

Difficulty should come from complex ideas, not complex sentences. When prose is hard to parse, the writer has shifted work to the reader. Expression techniques identify where this happens and provide repair strategies.

Most common failures — watch for these before loading technique files:

  • NEVER let abstractions float without grounding — load illustrate (which includes grounding moves) to anchor claims in concrete instances.
  • NEVER preserve passive voice because it sounds formal — passives hide actors, which hides responsibility.
  • NEVER conflate confidence with certainty — absolute language on inferred content creates dishonest prose.
  • NEVER allow AI diction patterns to survive: "it's worth noting", "it's important to", "in conclusion" — these are register failures, not stylistic choices.
  • NEVER fix word-level issues before structural ones — hedges and weak verbs on misordered prose are noise, not signal.
  • NEVER treat all hedging as weakness — hedging split (calibration vs. cushioning) is mandatory.

Default triage path: When unsure where to start, load tradition first, then check form-function alignment, then structural clarity — most prose problems are sequence problems. The default path is: tradition (load first) → fit → clarify → activate → strengthen → signal-confidence.

Diagnostic Questions

When improving prose ad-hoc, interrogate it with diagnostic questions. Each question that surfaces issues points to its technique file for detailed repair instructions.

Purpose

Diagnostic QuestionLoad Technique
Does the form match the function? Where does the shape undermine the intent?@./references/fit.md
What does this audience already know? Where does depth or convention mismatch?@./references/calibrate.md
What unfamiliar concepts appear without anchoring? What analogies could bridge?@./references/bridge.md

Clarity

Diagnostic QuestionLoad Technique
Can a reader follow the argument's logic — within this tradition? Where do concepts appear before their foundations?@./references/clarify.md
Are actors visible or hidden? Where does passive voice obscure who does what?@./references/activate.md
Does language commit or cushion? Where do hedges, weak verbs, or filler phrases weaken assertion? (Apply hedging split first)@./references/strengthen.md
Do abstractions have grounding? What claims need concrete examples? What witnessed detail could replace category descriptions?@./references/illustrate.md

Integrity

Diagnostic QuestionLoad Technique
Does certainty match evidence? Where does absolute language describe inferred content?@./references/signal-confidence.md
Where does this apply, and where does it break down? What boundary conditions go unstated?@./references/bound-scope.md
What must be true for this reasoning to hold? What premises operate silently?@./references/surface-assumptions.md

Shape

Diagnostic QuestionLoad Technique
What tension drives this forward? Is there shape, or just sequence?@./references/arc.md
Is the ordering purposeful? Where does the sequence feel arbitrary?@./references/arrange.md
Do sentences cluster at similar lengths? Where does rhythm flatten?@./references/rhythm.md
Does this sound like one person wrote it? Does that person speak from a stated place?@./references/voice.md
Does each word carry its weight? Does the emotional register fit the purpose?@./references/register.md

Depth

Diagnostic QuestionLoad Technique
What follows from these claims that goes unsaid? What implications lurk?@./references/extract-implications.md
Where does feedback conflate distinct concerns? What dimensions hide?@./references/dimensionalize.md
Does this prose close inquiry or open it? Where could questions invite exploration?@./references/pose-questions.md

Loading Techniques

To apply a technique, load its reference file:

@./references/{name}.md

CRITICAL: A technique reference file MUST be loaded before claiming to apply that technique. Reading the technique name in this skill file does NOT constitute applying the technique.

Technique Library

16 techniques in ./references/.

CategoryTechniques
Purposefit, calibrate, bridge
Clarityclarify, activate, strengthen, illustrate
Integritysignal-confidence, bound-scope, surface-assumptions
Shapearc, arrange, rhythm, voice, register
Depthextract-implications, dimensionalize, pose-questions
Meta-lenstradition (load before any other technique)

Load the full technique index for extended navigation:

@./references/index.md

Ad-Hoc Use

For applying techniques outside workflow routes:

  1. Load tradition lens — identify the rhetorical tradition before any diagnostic
  2. Check the omit gate — ask whether the content should exist before repairing it
  3. Interrogate the prose with diagnostic questions from the tables above
  4. Select technique(s) based on which questions surface issues
  5. Load technique reference file(s) for detailed instructions
  6. Follow the Instructions section of each loaded technique

NEVER

  • NEVER apply any diagnostics without first loading tradition — tradition meta-lens must precede all technique application.
  • NEVER treat all hedging as weakness — the hedging split is mandatory: distinguish epistemic calibration (preserve) from habitual cushioning (cut).
  • NEVER smooth productive friction without asking — irregular rhythm, unresolved tension, and deliberate register shifts are interrogated before normalized.
  • NEVER fill intentional gaps — gap-check precedes every addition: ask "is this gap doing rhetorical work?" before filling.
  • NEVER correct L1 grammatical patterns as errors — diagnose what the pattern encodes before treating it as a problem.
  • NEVER fix word-level issues before structural ones — hedges and weak verbs on misordered prose are noise, not signal.
  • NEVER default to Anglo-American conventions invisibly — name tradition-specific choices as tradition-specific, not universal.

Additional Resources

  • @./references/clarify-patterns.md — Detection heuristics and repair strategies
  • @./references/clarify-examples.md — Before/after transformations by issue type

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

40.6%
按下载量换算26

Claude

27.52%
按下载量换算17

Cursor

18.6%
按下载量换算12

Gemini CLI

10.38%
按下载量换算7

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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