Skill flagged — suspicious patterns detected

ClawHub Security flagged this skill as suspicious. Review the scan results before using.

Project Watcher

v1.0.0

项目规划与进度追踪。维护 roadmap,git commit 通知,远程部署感知,与飞书集成。

0· 86·0 current·0 all-time

Install

OpenClaw Prompt Flow

Install with OpenClaw

Best for remote or guided setup. Copy the exact prompt, then paste it into OpenClaw for altaircardinal/project-watcher.

Previewing Install & Setup.
Prompt PreviewInstall & Setup
Install the skill "Project Watcher" (altaircardinal/project-watcher) from ClawHub.
Skill page: https://clawhub.ai/altaircardinal/project-watcher
Keep the work scoped to this skill only.
After install, inspect the skill metadata and help me finish setup.
Use only the metadata you can verify from ClawHub; do not invent missing requirements.
Ask before making any broader environment changes.

Command Line

CLI Commands

Use the direct CLI path if you want to install manually and keep every step visible.

OpenClaw CLI

Bare skill slug

openclaw skills install project-watcher

ClawHub CLI

Package manager switcher

npx clawhub@latest install project-watcher
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Benign
high confidence
Purpose & Capability
The skill's name/description match its code and instructions: it manages a roadmap and sends git commit notifications to Feishu. Minor mismatch: registry metadata lists no required environment variables, but the code and SKILL.md expect Feishu credentials (app_id/app_secret) provided via env vars or configs/feishu.json.
Instruction Scope
SKILL.md instructs creating and updating ROADMAP.md and optionally installing a post-commit hook. The included post-commit hook and send_card.py only gather basic git metadata (branch, short hash, commit message, repo name) and POST to the official Feishu open-apis endpoint — scope matches the described notification purpose.
Install Mechanism
No install spec; this is instruction-only with small helper scripts included. Nothing is downloaded or executed automatically by an installer, so disk writes and execution are under user control (you must place the hook manually).
Credentials
The only required secrets in practice are Feishu app_id/app_secret (env vars or configs/feishu.json). The example projects.yaml references an SSH key path for remote_deploy (sensitive), but no code in this bundle reads or uses that SSH key. Be aware the skill encourages storing Feishu credentials in configs/feishu.json or env vars — the metadata not listing env requirements is an inconsistency to notice.
Persistence & Privilege
always is false and the skill does not request system-wide persistence or modify other skills. The post-commit hook runs only if you install it into a repository's .git/hooks; there is no automatic global installation.
Assessment
This skill is coherent with its description but review and control how you provide credentials and install hooks before using it: - send_card.py will send git commit metadata (repo, branch, short hash, commit message) to Feishu. You must supply FEISHU app_id/app_secret (via env vars FEISHU_APP_ID/FEISHU_APP_SECRET or by editing configs/feishu.json). Do not commit real secrets to repos. - The post-commit hook will run only if you put scripts/post-commit-hook.sh into .git/hooks/post-commit. Installing hooks is a manual action — don't copy hooks from untrusted code without review. - Example configs/projects.yaml references an SSH private key path. This repository does not use that key, but if you extend the skill to perform remote deploys, avoid placing private keys in project repos or shared config files. - Confirm the Feishu app has only the scopes you expect and that you trust the receive_id/open_id being used. Consider creating a dedicated Feishu app account with limited permissions for notifications. - If you need higher assurance, inspect send_card.py yourself; it uses the official Feishu endpoints and no obfuscated code was found. Given these checks, usage is reasonable. If the registry metadata should have declared required env vars, ask the publisher to update it before installing to avoid surprises.

Like a lobster shell, security has layers — review code before you run it.

latestvk973n58evcmfaj7brqja7gbzsn83snfm
86downloads
0stars
1versions
Updated 1mo ago
v1.0.0
MIT-0

SKILL.md — project-watcher

项目规划与进度追踪。维护 roadmap,让每个行动都能对应到规划条目。

核心原则

  1. 规划先行:新项目必须先有 roadmap 再开工,没有规划就开工是无效推进
  2. 条目可追踪:每个 roadmap 条目必须能回答"怎么算完成了"
  3. 进度有感知:推进项目后主动更新 roadmap,不等待商人询问
  4. 变更即更新:计划外的事项立即入 roadmap,不搁置

A. 项目初始化

触发:在非 workspace 根目录检测到 .git(首次进入项目目录工作)

流程

  1. 询问商人项目目标和技术栈
  2. 和商人一起拆解 roadmap(参考"拆解标准")
  3. 创建 project.mdROADMAP.md

B. Roadmap 条目设计规范

可接受的条目

- [ ] 分帧流水线(进度:0%)
  - 完成 criterion: 所有视频都能输出帧到 frames/
  - 依赖:无
  - 当前状态:等待素材

需要拆分的条目

  • ❌ "做偏好模型" → ✅ 拆成:文献调研、adapter 架构设计、训练数据准备、训练脚本、评估方案
  • ❌ "优化性能" → ✅ 拆成:profiling 定位瓶颈、制定优化方案、实施优化、验证效果

拆解由 agent 和商人共同完成,商人确认后再开工。

ROADMAP.md 格式

# [项目名称] Roadmap

## 当前版本 v0.x

### 进行中
- [ ] **功能名称**(进度:30%)
  - 完成标准:...
  - 依赖:功能A
  - 当前状态:正常推进

### 已完成
- [x] ~~已完成功能~~(v0.1)

## 下一步
- 调研方案(待确认)
- 功能B(依赖功能A完成后解锁)

## 技术债务
- 暂无

## 变更记录
- 2026-03-28:新增功能C(需求变更)

C. 项目推进流程(核心)

每次在项目目录推进工作时

触发条件:Agent 在项目目录执行 exec / spawn subagent / 调用工具后

Agent 行为

  1. 执行前:对照 roadmap,确认当前要做的是哪个条目
  2. 执行后:判断是否有进展 → 更新对应条目的进度或状态

进度更新判断

情况操作
条目完成[ ][x],进度标 100%,移至已完成区
部分进展更新进度 %,注明做了什么
遇到障碍更新"当前状态",告知商人卡点
发现新需求新增条目,标注"计划外"
依赖变更更新依赖关系,告知商人

更新后主动告知商人

✅ 已推进 [项目] — [条目名称]
  进度:30% → 45%
  做了什么:完成了人像过滤逻辑
  下一步:聚合函数设计

D. 变更同步

触发:项目中出现了 roadmap 没有记载的变化

变化类型处理
需求变更(商人提出)立即更新 roadmap,标注变更原因和日期
技术方案调整更新条目描述,说明原因
发现新任务新增条目,初始进度 0%
某条目取消移到"变更记录",注明取消原因

原则:变更发生后立即更新,不要攒到会话结束。


E. 进度回顾

触发:商人询问 / 每次会话开始时在项目目录工作

Agent 主动报告:

📋 [项目] 进度 — v0.3
  进行中:人像检测回溯(45%)
  下一步:聚合函数设计
  计划外:新增了视频质量评估需求(新增)
  阻塞:等待标注数据

F. 与其他模块的关系

  • post-commit hook:通知只负责通知,不触发 roadmap 更新
  • roadmap 更新:由 agent 在推进项目时自然完成
  • github skill:处理 GitHub Issues → 可作为 roadmap 条目来源

G. 配置

凭证和项目配置见对应文件:

  • configs/feishu.json — 飞书凭证(首次使用需填写自己的 app_id 和 app_secret
  • configs/projects.yaml — 项目路径和部署配置(根据需要添加项目)

Comments

Loading comments...