Token导航 LogoToken导航

Agent Skill · 研究检索

spec-workflow-mcp(spec 工作流 MCP)

spec-workflow-mcp 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。

spec-workflow-mcp 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为调用外部服务;许可证为 MIT-0,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 23,803。

总安装 4,176周安装 174GitHub Stars 公开资料未说明下载 1,392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
1.0.0

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:spec-workflow-mcp(spec 工作流 MCP)
来源仓库:https://github.com/bytesagain3/spec-workflow-mcp
安装命令:
openclaw skills install spec-workflow-mcp
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 spec-workflow-mcp

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。

通过 MCP 为 AI 辅助工作流程提供规范驱动的开发工具。在添加任务、规划迭代、跟踪完成情况、检查质量时使用。

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 spec-workflow-mcp;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

    它需要调用外部服务,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

spec-workflow-mcp 怎么安装?

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

spec-workflow-mcp 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。

spec-workflow-mcp 需要什么权限?

权限级别是调用外部服务。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。

spec-workflow-mcp 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。

spec-workflow-mcp 用的人多吗?

公开数据是累计安装 4,176 次,近一周安装 174 次,下载 1,392 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。

spec-workflow-mcp 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

  • 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。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 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。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 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
spec-workflow-mcp

spec 工作流 MCP · 本页

4,176—调用外部服务MIT-0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

name
Spec Workflow Mcp
description
Serve spec-driven dev tools via MCP for AI-assisted workflows. Use when adding tasks, planning iterations, tracking completion, reviewing quality.
version
1.0.0
license
GPL-3.0
runtime
python3

Spec Workflow MCP

Spec Workflow MCP v2.0.0 — a productivity toolkit for spec-driven development workflows served via the Model Context Protocol (MCP). Manage tasks, plan sprints, track progress, review deliverables, set reminders, prioritize work, and run weekly reviews — all from the command line with timestamped log entries.

Commands

The script (scripts/script.sh) exposes the following commands via a case dispatcher:

CommandDescription
add <input>Add a new spec/task entry. Without args, shows the 20 most recent add entries.
plan <input>Record a planning entry (sprint planning, iteration goals). Without args, lists recent plans.
track <input>Track progress on a task or deliverable. Without args, lists recent tracking entries.
review <input>Record a review note (code review, spec review). Without args, lists recent reviews.
streak <input>Log a streak entry (daily consistency tracking). Without args, lists recent streaks.
remind <input>Set a reminder or log a reminder note. Without args, lists recent reminders.
prioritize <input>Record a prioritization decision. Without args, lists recent prioritizations.
archive <input>Archive a completed item. Without args, lists recent archive entries.
tag <input>Tag or categorize an entry. Without args, lists recent tags.
timeline <input>Record a timeline/milestone entry. Without args, lists recent timeline entries.
report <input>Generate or log a report entry. Without args, lists recent reports.
weekly-review <input>Record a weekly review summary. Without args, lists recent weekly reviews.
statsShow summary statistics across all log files (entry counts per type, total, disk usage).
export <fmt>Export all data in json, csv, or txt format to $DATA_DIR/export.<fmt>.
search <term>Search all log files for a term (case-insensitive grep).
recentShow the 20 most recent lines from history.log.
statusHealth check — shows version, data directory, total entries, disk usage, last activity.
helpDisplay the full help/usage message.
versionPrint spec-workflow-mcp v2.0.0.

How Each Entry Command Works

  1. If called without arguments, it tails the last 20 lines of <command>.log.
  2. If called with arguments, it:

- Timestamps the input (YYYY-MM-DD HH:MM|<input>) - Appends it to $DATA_DIR/<command>.log - Prints confirmation with the current total count - Logs the action to history.log

Data Storage

All data is stored as plain-text log files under:

~/.local/share/spec-workflow-mcp/
├── add.log
├── plan.log
├── track.log
├── review.log
├── streak.log
├── remind.log
├── prioritize.log
├── archive.log
├── tag.log
├── timeline.log
├── report.log
├── weekly-review.log
└── history.log          # unified activity log

Each log line uses pipe-delimited format: YYYY-MM-DD HH:MM|<value>

The history.log uses: MM-DD HH:MM <command>: <value>

Requirements

  • Bash 4.0+ (uses local variables, set -euo pipefail)
  • coreutils: date, wc, du, tail, cat, basename, grep, sed
  • No external dependencies, API keys, or network access required
  • Works on Linux and macOS

When to Use

  1. Sprint planning — use plan to record iteration goals, then track to log progress against them throughout the sprint
  2. Task management — use add to capture new specs or tasks, prioritize to rank them, and archive when complete
  3. Daily standups — use streak to maintain consistency tracking and recent to review what happened yesterday
  4. Code/spec reviews — use review to log review notes and decisions for future reference
  5. Weekly retrospectives — use weekly-review to capture weekly summaries, then export to generate reports for stakeholders

Examples

Add a task and plan a sprint

# Add a new spec
bash scripts/script.sh add "API rate limiting — sliding window implementation"

# Plan the sprint
bash scripts/script.sh plan "Sprint 8: rate limiter, caching layer, monitoring"

Track progress and review

# Track completion
bash scripts/script.sh track "rate limiter: Redis backend done, unit tests passing"

# Log a review
bash scripts/script.sh review "rate limiter PR #87: approved, added integration tests"

Set reminders and manage priorities

# Set a reminder
bash scripts/script.sh remind "demo prep for Friday stakeholder meeting"

# Prioritize
bash scripts/script.sh prioritize "P1: cache invalidation edge case in multi-region"

Weekly review and export

# Record weekly summary
bash scripts/script.sh weekly-review "shipped rate limiter, started caching layer, blocked on Redis cluster config"

# Export all data as CSV
bash scripts/script.sh export csv

Search and view timeline

# Find all entries mentioning "cache"
bash scripts/script.sh search "cache"

# Log a milestone
bash scripts/script.sh timeline "v2.1 release candidate tagged"

# View overall statistics
bash scripts/script.sh stats

Configuration

Set the DATA_DIR variable (or modify it in the script) to change the storage directory. Default: ~/.local/share/spec-workflow-mcp/

Output

All commands print to stdout. Redirect to a file as needed:

bash scripts/script.sh weekly-review > retro-notes.txt

Powered by BytesAgain | bytesagain.com | hello@bytesagain.com

平台分布

OpenClaw

75.42%
按下载量换算1,050

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

调用外部服务

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。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