中文申请书五阶段研究与写作:课题准备、文献调研、方案制定、大纲规划、正文写作. An agent skill from dashbear-ai/grant-master.

MITAuto-check passed

Install Auto

skills CLI
$ npx skills add dashbear-ai/grant-master --skill auto -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install dashbear-ai/grant-master auto --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/dashbear-ai/grant-master.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/auto .claude/skills/auto && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
auto
GitHub stars
125
Token cost
~714 tokens
SKILL.md length
119 words
Files
2 (incl. references)
Skills in repo
2
Repo updated
First seen
Licence
MIT

At a glance

中文申请书五阶段研究与写作:课题准备、文献调研、方案制定、大纲规划、正文写作. An agent skill from dashbear-ai/grant-master.

  • SKILL.md covers 研究原则 and 与工作台合作
  • Calls python3

What it does

Auto is an agent skill from dashbear-ai/grant-master. 中文申请书五阶段研究与写作:课题准备、文献调研、方案制定、大纲规划、正文写作。 用户要求使用 Grant Master、推进申请书、恢复已有项目,或询问本 skill 如何使用时触发。 默认打开 Codex 内置工作台;研究在当前对话进行,网页展示资料并收集用户决策。

Its SKILL.md is about 710 tokens, which your agent loads only when the skill is triggered. The skill folder holds 2 other files, including reference files (for example `references/workbench-api.md`).

The repository describes itself as: 中文项目申请书全流程写作工具链 Grant-Master,支持课题理解、文献调研、论文精读、方案收敛、大纲规划、正文写作、审阅与 docx 输出。 The licence is MIT.

Example prompts

  • “/auto”

Requirements

  • Python 3

What it can do on your machine

Read from SKILL.md and the folder at commit 23c4948. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Auto loads about 714 tokens when it runs, and up to ~2.4k if it reads all its reference files. Until then it costs about 35 tokens; SKILL.md has 119 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~35
When it runs · the whole SKILL.md, loaded when a task matches
~714
With references · SKILL.md plus every file in references/, read only if the agent opens them
~2.4k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from dashbear-ai/grant-master at commit 23c4948, republished under its MIT licence (© dashbear-ai). 119 words, ~714 tokens.

Download SKILL.mdSave it as .claude/skills/auto/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
auto
description
中文申请书五阶段研究与写作:课题准备、文献调研、方案制定、大纲规划、正文写作。 用户要求使用 Grant Master、推进申请书、恢复已有项目,或询问本 skill 如何使用时触发。 默认打开 Codex 内置工作台;研究在当前对话进行,网页展示资料并收集用户决策。

Grant Master

直接启动工作台 · Linux/WSL:运行 python3 ../../workbench/launch.py。无需先请求 AI。

首次使用或恢复时明确告诉用户:主流程在 Codex 的 AI 对话框中进行,网页只作为报告展示、资料编辑和辅助交互。请配合 Codex 一起使用;AI 停止后,请回到此对话发送“继续”。

即使用户只是询问用法,也先运行插件的 workbench/launch.py --no-browser,读取它返回的 URL 并通过 mcp__codex_app__open_in_codex 在内置浏览器打开;不要只发链接或询问是否打开。用户明确不打开时尊重其选择。只问用法时介绍 Demo,不创建项目或启动研究。

研究原则

每次对话开始,hook 只检查小型修改索引;发现未读记录后,先运行 gm.py changes --project ID,按 more 逐页读到 false,理解并响应用户修改,再继续研究。没有 hook 提示时,恢复项目仍通过 context 核对 userRevision 与 changesReadRevision。读取回执仅表示已接收,不代表修改要求已经完成;把未完成工作说明留在研究报告或待办中,不通过手写状态绕过后端。

进入阶段时先读取 context 返回的 stageGuidance 中对应文件。这是网页“本阶段研究经验”展示的同一份长期共用指导;文献调研前运行 gm.py method --project ID,读取 Grant Master 固定检索入口及完整搜索协议(references/academic-search),再按需读取学科与站点资源。CLI 回执绑定入口和核心协议内容;修改后需重新读取。研究报告说明实际采用的方法步骤和证据范围,不能只说“已遵循”。方法只提供研究规则,其外部操作仍受用户授权与可用工具约束。

研究文件以工作台返回的项目目录为准,与 Codex 当前工作目录无关。用户新建项目默认位于 ~/.grant-master/<随机项目编号>/;位置可更改。原始资料、下载论文、候选稿和交付物都放入该项目,外部课题资料先导入副本。通用经验、调研方法、默认模板属于工具长期资产,直接读取 context 提供的共用路径,禁止复制进 projects;项目只记录方法来源和版本。只读取当前阶段需要的报告,需要核实具体论据时再追溯原件。

  • 课题准备:先区分用户已有事实、你的假设和待确认事项。把宽泛主题收敛成研究对象、问题、条件和边界;尚未调研时不声称发现文献空白。
  • 文献调研:使用 Grant Master 学术检索。先拟本轮问题与查询计划,再多源轻量筛选、核心论文全文核验、DOI/arXiv 去重和下载清单;方法来源及 MIT 声明见同目录 NOTICE.md。当前 AI 默认执行研究,只按需读相关参考,不强制启动 worker。检索报告与结构化清单通过 gm 发布到本项目 literature。每篇阅读报告解释研究问题、方法、关键结果、局限和启发,最新视角比较证据。
  • 方案制定:围绕一条可验证主线,讲清问题为何值得做、已有方法为何不足、拟议方法为何可能有效,以及如何证伪。指标、资源和实验应匹配申请人条件,未确认条件不能写成已有基础。
  • 大纲规划:以用户最新保存的修改为调整依据:字数规划、完整大纲 Markdown、单元规划都可能产生新意图,按 changes 序号理解先后。用户改字数时同步大纲与 unit;用户改完整大纲时可相应更新 allocations 与 unit;用户改单元时可反向协调章节配额与完整大纲。三者地位相同,不用某页的旧值压回最新修改。若修改间确有歧义,通过待办澄清。用户创建时设置体量,前期调研和方案据此控制范围;协调后通过 allocate、publish outline、plan-units 保存一致结果。完整大纲每个主要部分使用同级 Markdown 标题,并在其下标注 目标字数:N 字,发布时传当前 outlineStatus.revision 为 outline_revision。用户修改后的三个视图可暂时不同,AI 恢复后负责协调;不把网页不一致提示作为下一步入口。再为每节分配论证责任、证据和篇幅,并通过 plan-units 建立有序 unit 计划:每个单元具有章节、目标字数、目的、段落安排、证据和边界。章节引言、标题与叶子章节都要有明确归属,确认大纲前完成单元预算校验。大纲必须能读出完整推理链,避免重复背景、同一创新点多处改写以及空泛标题。
  • 正文写作:依据已确认的 unit 计划逐单元写作,按 context 状态续写和返修;发布带单元计划版本,不重写已完成单元。全部完成后按计划 assemble 成完整 Markdown。审阅意见定位到单元,修订后重新组装;全文有人工编辑时先合并候选并确认覆盖版本。术语统一,首次解释缩写;事实、计划和预期效果分开;引用能追溯到项目资料。审阅、修订、组装与 DOCX 导出都属于本阶段。段落、标题、表格之间保留空行以保证导出排版。

与工作台合作

使用 workbench/gm.py 的 context 读取资产、校验状态和未处理回答。currentStage 表示最早需要重新验证的位置,不代表用户退回了研究阶段;先响应最新修改,不机械重跑已有研究。继续时先处理已提交回答和拒绝,处理完成后确认;已有问题继续等待原问题,不重复投递。用户只要求当前一步时止于该范围,要求完整流程时持续推进至交付。

内容交付及等待的具体命令见 工作台调用,按需读取。你只提供研究内容、判断和待确认的问题;项目状态、资产索引、版本和事件记录由后端创建和维护,不手写状态文件。

需要用户参与时向统一待办中心投递解释、背景、选项和自由输入,等待真实回答;用户拒绝不视为同意。网页保存动作不启动模型。回答在停止期间也会保留,下次恢复后接收。不能把“请去网页点击下一步”当作本次研究已完成。

© dashbear-ai, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (references) in skills/auto of dashbear-ai/grant-master.

  • SKILL.md
  • references/workbench-api.md

Open the folder on GitHubat commit 23c4948

Compare with similar skills

Auto next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Auto compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Auto this skilldashbear-ai/grant-master125—~714Automated safety check: PassMIT
Grantsalirezarezvani/claude-skills28k—~3.7kAutomated safety check: PassMIT
Research GrantsK-Dense-AI/scientific-agent-skills48k1 repos~4.3kAutomated safety check: NotesMIT
Grantsborghei/Claude-Skills886—~1.8kAutomated safety check: PassMIT
Pp Grantsmvanhorn/printing-press-library2.1k—~1.5kAutomated safety check: NotesApache-2.0
Grant Proposalmohitagw15856/pm-claude-skills1.4k—~1.1kAutomated safety check: PassMIT

Similar skills

  • Grants

    alirezarezvani/claude-skills

    NIH grant research skill for clinical researchers. An agent skill from alirezarezvani/claude-skills.

    28k GitHub stars~3.7k tokensUpdated 1 mo ago
    Documents & OfficeAuto-check passed
  • Research Grants

    K-Dense-AI/scientific-agent-skills

    Supports research proposal preparation and review for NSF, NIH, DOE, DARPA, and Taiwan NSTC, including opportunity-specific requirements, aims, review criteria, budgets, broader impacts, forms, and…

    48k GitHub starsUsed in 1 repo~4.3k tokens
    Research & ScienceAuto-check: notes
  • Grants

    borghei/Claude-Skills

    Grant writing and proposal architecture: funder fit, proposal structure, budget design, and success-factor scoring.

    886 GitHub stars~1.8k tokensUpdated 2 days ago
    Research & ScienceAuto-check passed
  • Pp Grants

    mvanhorn/printing-press-library

    Find open US federal research funding and benchmark award sizes — Grants.gov open opportunities plus awarded NIH RePORTER and NSF grants, keyless.

    2.1k GitHub stars~1.5k tokensUpdated yesterday
    Research & ScienceAuto-check: notes
  • Grant Proposal

    mohitagw15856/pm-claude-skills

    Write a structured grant proposal or funding application for any grant type.

    1.4k GitHub stars~1.1k tokensUpdated yesterday
    Research & ScienceAuto-check passed
  • Review Grant

    marin-community/marin

    Review an explicitly identified marin-iac grant PR that edits IAM data or a deploy-target module, confirm its decrypted principals and roles, then apply only the confirmed grant.

    3.9k GitHub stars~1.1k tokensUpdated today
    DevOps & CloudAuto-check passed

More from dashbear-ai/grant-master

  • Academic Search

    dashbear-ai/grant-master

    Grant Master 文献调研阶段的多源检索、两遍筛选、开放全文获取和元数据合并方法. An agent skill from dashbear-ai/grant-master.

    125 GitHub stars~758 tokensUpdated 1 mo ago
    Auto-check passed

Questions about Auto

What does Auto do?

中文申请书五阶段研究与写作:课题准备、文献调研、方案制定、大纲规划、正文写作. An agent skill from dashbear-ai/grant-master. Auto is an agent skill from dashbear-ai/grant-master.

How do I install Auto in Claude Code?

Run `npx skills add dashbear-ai/grant-master --skill auto -a claude-code`. Or copy the skill folder (skills/auto in dashbear-ai/grant-master) into .claude/skills/auto in your project. Claude Code loads it when a task matches its description.

How do I install Auto in Codex?

Run `npx skills add dashbear-ai/grant-master --skill auto -a codex`. Or copy the skill folder (skills/auto in dashbear-ai/grant-master) into .agents/skills/auto in your project. Codex loads it when a task matches its description.

Can I use Auto in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add dashbear-ai/grant-master --skill auto -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/auto, .gemini/skills/auto, .github/skills/auto and .opencode/skills/auto in your project.

What does Auto need to run?

Going by SKILL.md and its folder, Auto needs the command-line tools its instructions call (python3). Our summary lists: Python 3.

Does Auto access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is Auto safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Auto use?

Auto is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Auto use?

About 714 tokens (SKILL.md is roughly 2.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 1.7k tokens, read only when the agent opens those files.

What are the alternatives to Auto?

Skills that share tags, products or a category with Auto: Grants (alirezarezvani/claude-skills, 28k stars), Research Grants (K-Dense-AI/scientific-agent-skills, 48k stars), Grants (borghei/Claude-Skills, 886 stars) and Pp Grants (mvanhorn/printing-press-library, 2.1k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Auto?

dashbear-ai (a GitHub organization) maintains it in dashbear-ai/grant-master, which has 125 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on September 9, 2026.

Source: dashbear-ai/grant-master on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.