Token导航 LogoToken导航

Agent Skill · 研究检索

dictation-instructions(听写指令)

用于辅助提示词、系统指令、Agent 行为约束和工作流模板的整理。它适合让 Agent 规范任务边界、统一输出格式、拆分操作步骤或优化提示词可复用性。使用时需要保留真实业务约束,不要把示例当硬规则;涉及自动执行、外部工具或高风险操作时,应在提示词中明确确认步骤、权限边界和失败处理方式。

dictation-instructions 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 7,164。

总安装 19,760周安装 786GitHub Stars 4,412下载 10,287核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:dictation-instructions(听写指令)
来源仓库:https://github.com/github/gh-aw
仓库路径:skills/dictation-instructions
安装命令:
npx skills add https://github.com/github/gh-aw --skill 'Dictation Instructions'
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/github/gh-aw --skill 'Dictation Instructions'

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

用于规范提示词结构和 Agent 行为约束条件。帮助拆分操作步骤、统一输出格式和优化模板复用性。包含 GitHub Agentic Workflows 项目专用术语校正规则。需保留真实业务约束,避免将示例当作硬性规则。涉及高风险操作时应明确确认步骤与权限边界。

dictation-instructions 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/dictation-instructions

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/github/gh-aw --skill 'Dictation Instructions';命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

dictation-instructions 怎么安装?

在终端执行 npx skills add https://github.com/github/gh-aw --skill 'Dictation Instructions'。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

dictation-instructions 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

dictation-instructions 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

dictation-instructions 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

dictation-instructions 用的人多吗?

公开数据是累计安装 19,760 次,近一周安装 786 次,下载 10,287 次,GitHub 4,412 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

dictation-instructions 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 4,412 星,维护活跃度相对可查。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
dictation-instructions

听写指令 · 本页

19,7604,412接触敏感数据未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Dictation Instructions

Technical Context

GitHub Agentic Workflows (gh-aw) is a Go-based GitHub CLI extension for writing AI-powered workflows in natural language using markdown files that compile to GitHub Actions YAML.

Project Glossary

The following project-specific technical terms should be corrected when encountered in speech-to-text input:.github.github/aw.github/workflows.lock.yml.md @copilot ACTIONS_STEP_DEBUG ANTHROPIC_API_KEY CLAUDE_CODE_OAUTH_TOKEN CODEX_API_KEY COPILOT_GITHUB_TOKEN DEBUG FUZZY:BI-WEEKLY FUZZY:DAILY FUZZY:WEEKLY GEMINI_API_KEY GH_AW_AGENT_TOKEN GH_AW_ALLOWED_DOMAINS GH_AW_GITHUB_TOKEN GH_AW_PROMPT GH_AW_SAFE_OUTPUTS GH_AW_VERSION GH_AW_WORKFLOW_ID GH_TOKEN GITHUB_TOKEN OPENAI_API_KEY RUNNER_TEMP SARIF activation activation-job agent-job agentic agentic-workflows allow-workflows allowed allowed-domains allowed-files allowed-labels allowed-repos allowlist api.github.com artifact assign-to-agent assign-to-copilot audit audit-workflows auto-merge automation bash branch branch-name build bun cache cache-key cache-memory checkout checks claude code-scanning codex coding-agent comment compile compile-workflow compiled compiler concurrency concurrency-group config contents create-discussion create-issue create-pull-request cron custom custom-agent daily debug default-branch defaults deno dependabot description dispatch-workflow discussions docker documentation dotnet draft edit engine engine-config engines environment environment-variables events expires fail-fast fallback-as-issue feature-flag features frontmatter fuzzy fuzzy-schedule gateway gemini gh-aw git-branch git-commit github github-actions github-app github-token github.repository github.run_id github/gh-aw hourly id-token imports input inputs inspect installation integration issue issue-ops issue_comment issueops issues javascript job-discriminator jobs json json-schema label label-ops labelops labels latest lockfile logs main markdown max-continuations max-turns mcp mcp-gateway mcp-inspect mcp-list mcp-registry mcp-scripts mcp-server mcp-servers merge metadata milestone min-integrity mode model needs.activation network network.allowed network.firewall node noop on-demand org organization outputs override owner package parallel parsing permissions phase pip pipeline playwright prompt prompt-injection protected-files pull-requests pull_request pull_request_target python python3 recompile registry release repo repo-memory report-summary repository repository_dispatch review runs-on runtime runtimes safe-inputs safe-outputs sandbox sanitized schedule scheduled schema scripts search secrets security session setup shared shared-workflow shell slack staged staged-mode stale steps.sanitized.outputs.body steps.sanitized.outputs.text steps.sanitized.outputs.title timeout timeout-minutes token-weights toolsets trigger triggers trusted trusted-users ubuntu ubuntu-latest update-issue update-pull-request validate validation version version-bump vulnerability-scan web-fetch web-search webhook weekly workflow workflow-dispatch workflow-run workflow-status workflow_call workflow_dispatch workflow_run workflows workspace write-all yaml zizmor

Fix Speech-to-Text Errors

When fixing dictated text, correct these common misrecognitions:

GitHub and Git Terms

  • "get hub" → github
  • "git lab" → gitlab
  • "get actions" → github-actions
  • "pull request" → pull-request (when used as compound modifier)
  • "issue ops" → issueops
  • "label ops" → labelops
  • "chat ops" → chatops
  • "multi repo ops" → multirepoops
  • "project ops" → projectops
  • "data ops" → data-ops
  • "dispatch ops" → dispatch-ops
  • "daily ops" → daily-ops

Workflow Configuration

  • "front matter" → frontmatter
  • "safe outputs" → safe-outputs (in configuration context)
  • "safe inputs" → safe-inputs (in configuration context)
  • "lock file" →.lock.yml or lockfile (depending on context)
  • "tool sets" → toolsets
  • "M.C.P." or "M C P" → MCP
  • "repo memory" → repo-memory (in configuration context)
  • "cache memory" → cache-memory (in configuration context)
  • "work flow" → workflow
  • "timeout minutes" → timeout-minutes
  • "runs on" → runs-on
  • "min integrity" → min-integrity (in configuration context)
  • "mcp gateway" → mcp-gateway
  • "mcp scripts" → mcp-scripts
  • "staged mode" → staged-mode
  • "token weights" → token-weights
  • "effective tokens" → effective-tokens

AI Engines

  • "co-pilot" → @copilot
  • "code x" → codex
  • "cloud" → claude (when referring to the AI engine)
  • "gem ini" → gemini (when referring to the AI engine)
  • "serena" → serena (code intelligence MCP server)

Commands and Operations

  • "G.H. A.W." → gh-aw or gh aw (depending on context)
  • "re-compile" → recompile
  • "work flow dispatch" → workflow_dispatch
  • "action lint" → actionlint
  • "ziz more" → zizmor
  • "poo teen" → poutine
  • "queue M.D." → qmd

File Formats and Extensions

  • "dot M.D." →.md
  • "dot Y.A.M.L." or "dot Y M L" →.yaml or.yml
  • "dot lock dot Y M L" →.lock.yml
  • "jason" → JSON (when referring to format)
  • "wasm" → WebAssembly or wasm (depending on context)

Technical Patterns

  • "A.P.I." → API
  • "U.R.L." → URL
  • "H.T.T.P." → HTTP
  • "H.T.T.P.S." → HTTPS
  • "S.H.A." → SHA
  • "C.I." → CI
  • "G.H." → GH (when referring to GitHub CLI)
  • "Y.A.M.L." → YAML
  • "O.I.D.C." → OIDC
  • "S.A.R.I.F." → SARIF

Hyphenation Rules

Use hyphens for compound modifiers:

  • "safe outputs" → safe-outputs
  • "safe inputs" → safe-inputs
  • "cache memory" → cache-memory
  • "timeout minutes" → timeout-minutes
  • "cross repository" → cross-repository
  • "pull request" → pull-request (when used as adjective)
  • "mcp gateway" → mcp-gateway
  • "mcp scripts" → mcp-scripts
  • "token weights" → token-weights

Environment Variables

Capitalize fully: GITHUB_TOKEN, GH_TOKEN, COPILOT_GITHUB_TOKEN, GH_AW_GITHUB_TOKEN, ANTHROPIC_API_KEY, OPENAI_API_KEY, GEMINI_API_KEY, CLAUDE_CODE_OAUTH_TOKEN, CODEX_API_KEY

Common Ambiguities

  • "their/there/they're" → use context to determine correct spelling
  • "its/it's" → its (possessive), it's (it is)
  • "your/you're" → your (possessive), you're (you are)

Clean Up and Improve Text

Remove filler words and improve clarity:

Remove These Filler Words

  • humm, um, uh, uhh, umm
  • you know, like, basically, actually, literally
  • kind of, sort of, I mean, I think
  • right?, okay?, so yeah, well

Improve Clarity

  1. Remove redundant phrases:

- "in order to" → "to" - "at this point in time" → "now" - "due to the fact that" → "because" - "in the event that" → "if"

  1. Make text more concise:

- Remove unnecessary qualifiers (very, really, quite) - Use active voice instead of passive voice - Replace wordy phrases with simpler alternatives

  1. Maintain technical accuracy:

- Keep all technical terms from the glossary - Preserve code examples and commands exactly - Don't simplify technical concepts

Guidelines

You do not have enough background information to plan or provide code examples.

  • do NOT generate code examples
  • do NOT plan steps
  • focus on fixing speech-to-text errors and improving text quality
  • remove filler words (humm, you know, um, uh, like, basically, actually, etc.)
  • improve clarity and make text more professional
  • maintain the user's intended meaning

平台分布

Codex

35.5%
按下载量换算3,652

Claude

30.63%
按下载量换算3,151

Cursor

20.21%
按下载量换算2,079

Gemini CLI

10.95%
按下载量换算1,126

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills