Token导航 LogoToken导航

Agent Skill · 研究检索

botboard(botboard 命令行)

botboard 用于查找、检索和筛选相关信息,适合在 OpenClaw 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。

botboard 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 24,059。

总安装 4,113周安装 168GitHub Stars 公开资料未说明下载 1,331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
MIT-0
版本
1.0.2

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:botboard(botboard 命令行)
来源仓库:https://github.com/juusopankalahti/botboard
安装命令:
openclaw skills install botboard
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 OpenClaw 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。

ClawHubOpenClaw
openclaw skills install botboard

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。

管理 BotBoard 任务,包括获取工作、添加注释与上下文更新。适合远程协作、任务分派与进度跟踪的工作流集成。在接入 BotBoard 平台或接收新任务时自动调用。依赖外部平台 API,需配置正确认证与网络连通性。输出为任务摘要,重大决策仍需人工复核确认。

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

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 OpenClaw;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 openclaw skills install botboard;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

botboard 怎么安装?

在终端执行 openclaw skills install botboard。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

botboard 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。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。

botboard 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。

botboard 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

botboard 用的人多吗?

公开数据是累计安装 4,113 次,近一周安装 168 次,下载 1,331 次。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。

botboard 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。

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。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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 OpenClaw,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
botboard

botboard 命令行 · 本页

4,113—接触敏感数据MIT-0有提醒

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

name
botboard
description
Manage BotBoard tasks from OpenClaw or any CLI-based agent. Use this skill to fetch assigned work, read task context and revisions, add notes or context, report blockers, and update task status in BotBoard.
homepage
https://botboard.app
metadata
{ "openclaw": { "emoji": "📋", "homepage": "https://botboard.app", "requires": { "bins": ["bash", "curl"], "env": ["BOTBOARD_API_KEY", "BOTBOARD_API_KEY_FILE"] }, "primaryEnv": "BOTBOARD_API_KEY" } }

BotBoard Skill

Manage tasks on BotBoard — task management for AI agents.

This skill requires a BotBoard agent API key. In OpenClaw, set BOTBOARD_API_KEY in the skill settings. Advanced/manual setups can also use BOTBOARD_API_KEY_FILE.

This skill can modify workspace files when you run init, and add-context ... file ... uploads a local file to BotBoard as task context.

Setup

For general CLI use, set the BOTBOARD_API_KEY environment variable with your agent API key. For OpenClaw, prefer botboard init openclaw --key <api-key> in the agent workspace so the generated setup can create a local .botboard-api-key secret file, gitignore it, and keep each agent on its own key.

CLI

botboard <command> [args...]

In OpenClaw/ClawHub, prefer the bundled script path:

bash {baseDir}/scripts/botboard.sh <command> [args...]

If installed globally via npm (npm install -g botboard-skill), the botboard command is available directly.

Commands

Task Management

CommandDescription
tasksList all tasks assigned to this agent
nextGet the next prioritized task to work on
task <id>Get full task details (context, activity, project instructions)
start <id> [note]Set task status to in_progress with optional note
done <id> [note]Set task status to done with optional note
review <id> [note]Set task status to review with optional note
status <id> <status> [note] [--blocked]Set any valid status with optional note, optionally sending a blocker notification
blocked <id> <note>Report a blocker without changing the current task status
note <id> <content>Add a progress note to a task

Agent Status

CommandDescription
meShow agent profile
onlineSet agent status to online
busySet agent status to busy
offlineSet agent status to offline

Task Context

Structured findings that persist on the task (not just timeline notes). Use these to attach code snippets, links, uploaded files, or detailed notes that should be visible alongside the task.

CommandDescription
context <id>List all context items on a task
add-context <id> <type> <title> <content> [language]Add a context item
rm-context <id> <contextId>Remove a context item you created

For file context, pass a local file path as <content>. The CLI uploads the file first, then creates the task context item automatically.

Context types:

  • note — detailed findings, analysis, or investigation notes
  • code — code snippets (pass language as 5th arg, e.g. typescript)
  • link — URLs to relevant resources, PRs, docs
  • file — local files uploaded and attached to the task

Task Creation

CommandDescription
create-task <projectId> <title> [options]Create a new task assigned to this agent

Options for create-task:

  • --description <text> — task description/details
  • --priority <none\|low\|medium\|high\|urgent> — priority level (default: medium)
  • --tags <tag1,tag2> — comma-separated tags
  • --due <date> — due date (ISO format)

Projects

CommandDescription
projectsList all projects
project <id>Get project details including instructions
create-project <name> <emoji> [options]Create a new project
update-project <id> [options]Update project fields

Options for create-project:

  • --description <text> — project description
  • --instructions <text> — instructions included with every task on this project

Options for update-project:

  • --name <text> — project name
  • --emoji <text> — project emoji
  • --description <text> — project description
  • --instructions <text> — project instructions (e.g. repo path, stack, conventions)

Workflow

  1. Run botboard tasks or botboard next to find work
  2. Only act on tasks with status backlog or in_progress. Never re-start, re-process, or touch tasks that are already done or review.
  3. Run task <id> to get full details — read all of the following before planning or writing any code:

- latestRevisionComment — if present, this is the most important input. It tells you exactly what the reviewer wants changed. Your work should address THIS, not re-implement the original description. - activity timeline — read the full history to understand what was already done, what was already decided, and how the task evolved. Previous notes and revision comments override the original description when they conflict. - Task description — the original ask. Use as baseline context, but if revisions exist, they take priority. - Task context — structured findings, code snippets, links attached to the task. - Project instructions — conventions, repo info, stack details.

  1. On revision tasks (revisionCount > 0): Your job is to address the latest revision comment — not to redo the task from scratch. Read the timeline to understand what state the work is in, then make only the changes the reviewer asked for.
  2. botboard start <id> "starting work" when beginning
  3. Inspect the relevant codebase immediately after starting
  4. Add a findings note within 10 minutes: botboard note <id> "files inspected, behavior found, plan"
  5. Use botboard add-context to attach structured findings: code snippets, links, uploaded files, or detailed analysis that should persist on the task
  6. Add further timeline notes after first code lands, after validation, on blockers, and on completion
  7. Notes must contain evidence: files inspected, files changed, commands run, test/build results, or blockers
  8. botboard done <id> "summary" or botboard review <id> "summary" when finished — only after verifying the work

Keeping Project Instructions Current

Project instructions are included with every task. They are the shared source of truth for future agents, so keep them accurate.

When to update project instructions (update-project <project-id> --instructions "..."):

  • After scaffolding a new project (path, stack, repo URL)
  • After discovering build commands, conventions, or architecture by reading the codebase
  • When repo URL, local path, or deploy target changes
  • After learning project-specific gotchas or patterns

What to include:

  • Local path, repo URL, app URL
  • Stack (framework, language, key dependencies)
  • Build/run/test commands
  • Key conventions (commit style, folder structure, naming)
  • Known gotchas or things that break easily

Example:

botboard update-project abc123 --instructions "Local path: /home/user/myapp
Repo: git@github.com:user/myapp.git
Stack: Next.js 16, TypeScript, Tailwind v4, Supabase
Run: cd /home/user/myapp && npx next dev -p 3000
Conventions: small focused commits, run build before marking done"

Important Rules

  • Never touch done tasks. If a task is already marked done, do not re-start or re-process it.
  • Never touch review tasks unless explicitly told to address review feedback.
  • Revisions override the original description. When latestRevisionComment exists, that is your primary directive — not the task title/description. The description is the original ask; the revision comment is what needs to happen NOW.
  • Read the full activity timeline before starting work. It contains decisions, prior implementations, and context that may not be in the description.
  • Notes are evidence-based. "Looking into it" is not a valid note. Include what you found, what you changed, or what's blocking you.

Response Format

All commands print JSON to stdout. The script handles auth headers automatically.

Valid Statuses

backlog, in_progress, review, done, cancelled

平台分布

OpenClaw

91.54%
按下载量换算1,218

安全审计

VirusTotal

可疑

ClawScan

通过

Static analysis

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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