← 返回 Skills 市场

Clawwork

Name: Clawwork
Author: felipetruman

作者 felipetruman · GitHub ↗ · v1.0.0

cross-platform ⚠ suspicious

557

总下载

当前安装

版本数

在 OpenClaw 中安装

/install clawwork

功能描述

Execute tarefas profissionais via ClawWork - transforme Zero em um coworker de IA economicamente viável. Use quando precisar executar trabalhos complexos, ge...

使用说明 (SKILL.md)

ClawWork Skill

Integração com ClawWork para execução de tarefas profissionais com tracking econômico.

O que é

ClawWork transforma o assistente em um "coworker de IA" que:

Executa tarefas profissionais reais (220 tarefas em 44 setores)
Paga por tokens utilizados
Ganha "dinheiro" ao completar tarefas com qualidade
Mantém balance econômico (sobrevivência)

Comandos

CLI

# Status dos agentes
skill clawwork status

# Comparar modelos
skill clawwork compare

# Executar tarefa (requer E2B_API_KEY)
skill clawwork run -t "Criar análise de mercado"
skill clawwork run -t "Gerar plano de marketing" -m kimi-coding/k2p5

Uso Direto

# Via script
/home/freedom/.openclaw/workspace/skills/clawwork/clawwork.sh status

# Via Python
python /home/freedom/.openclaw/workspace/skills/clawwork/cli.py status

Setup

1. Configurar API Keys

Edite ~/.openclaw/workspace/ClawWork/.env:

# OpenRouter (já configurado ✅)
OPENAI_API_KEY=sk-or-v1-xxx
OPENAI_API_BASE=https://openrouter.ai/api/v1

# E2B - necessário para execução de código
# Obtenha gratuitamente em: https://e2b.dev/
E2B_API_KEY=e2b_xxx

2. Iniciar Dashboard (opcional)

cd ~/.openclaw/workspace/ClawWork
./start_dashboard.sh

Acesse: http://localhost:3000

3. Testar

skill clawwork status

Funcionamento

Usuário: /clawwork "Criar análise de mercado para SaaS B2B"

Zero → Classifica tarefa → [Technology / Software Engineer]
     → Define valor → [$50 baseado em salário BLS]
     → Executa via GLM-4.7 + ferramentas
     → Avalia qualidade → [GPT-4o scoring]
     → Calcula economics:
       💸 Custo tokens: $2.50
       💵 Pagamento: $45.00 (qualidade 90%)
       📊 Lucro: $42.50
     → Retorna resultado + métricas

Dados Existentes

Já temos testes anteriores:

GLM-4.7: 157 dias de logs
Kimi K2.5: 220 dias de logs
Qwen3 Max: 220 dias de logs

Localização

Componente	Path
ClawWork	`~/.openclaw/workspace/ClawWork/`
Skill	`~/.openclaw/workspace/skills/clawwork/`
Config	`~/.openclaw/workspace/ClawWork/.env`
Dados	`~/.openclaw/workspace/ClawWork/livebench/data/`

Mais Informações

Repositório: https://github.com/HKUDS/ClawWork
Dataset: GDPVal
Leaderboard: https://hkuds.github.io/ClawWork/

安全使用建议

This skill largely does what it says (wraps a ClawWork project), but there are important inconsistencies and risks you should consider before installing: - Secrets: The skill expects OPENAI_API_KEY / OPENAI_API_BASE and an E2B_API_KEY stored in ~/.openclaw/workspace/ClawWork/.env, but the registry metadata does not declare these. Do not place high-privilege keys there until you trust the code. Prefer scoped/test keys. - Hardcoded paths: Files use absolute paths under /home/freedom and expect a venv at CLAWWORK_DIR/venv. If your environment differs the code may fail or read unexpected files. Review and adapt paths before running. - Unreviewed imports: The CLI imports agent.live_agent from the ClawWork project directory (not included in these files). That code will run when you execute tasks — inspect the ClawWork project (the referenced GitHub repo) before executing to ensure it doesn't perform unwanted network, filesystem, or credential operations. - Run in isolation first: If you want to try it, run it in a disposable VM or container and use limited-scope API keys. Verify what network endpoints are contacted (OpenRouter/E2B and any other endpoints the ClawWork project contacts). - Metadata fix: Ask the skill author/maintainer to update the registry metadata to declare required env vars (OPENAI_API_KEY, OPENAI_API_BASE, E2B_API_KEY) and to avoid hardcoded user paths. Given the mismatches (undeclared secrets + code that will import external project modules and use them), treat this skill as suspicious until you can review the full ClawWork project and confirm the exact runtime behavior.

功能分析

Type: OpenClaw Skill Name: clawwork Version: 1.0.0 The skill is suspicious due to a critical prompt injection vulnerability that can lead to Remote Code Execution (RCE). The `cli.py` script directly embeds user-provided `task_description` into a configuration passed to an AI agent (`LiveAgent`). The `SKILL.md` explicitly states that an `E2B_API_KEY` is 'necessário para execução de código' (necessary for code execution), indicating the agent's capability to execute code. This allows an attacker to craft a malicious `task_description` to instruct the underlying AI (via E2B) to execute arbitrary commands on the host system, posing a significant security risk.

能力评估

ℹ Purpose & Capability

The name/description (ClawWork integration to run paid AI tasks) align with the included code (CLI, quick task, shell wrapper). However the registry metadata lists no required environment variables or credentials while the SKILL.md and code clearly expect OpenRouter/OpenAI and E2B API keys in ~/.openclaw/workspace/ClawWork/.env. Also the code uses absolute paths under /home/freedom which may not match a different installation environment.

⚠ Instruction Scope

SKILL.md and the Python code instruct the agent/user to read and edit ~/.openclaw/workspace/ClawWork/.env, start a local dashboard, and run the CLI. The CLI writes a temporary JSON config under the ClawWork workspace, enumerates local agent data, and imports project modules (agent.live_agent) from the ClawWork codebase. Importing and running those modules may execute arbitrary logic not shown in the skill bundle. The instructions also reference contacting external services (OpenRouter/OpenAI and E2B) — which is coherent with the function but means secrets will be used externally.

ℹ Install Mechanism

There is no install spec (instruction-only skill with included code files). That lowers installer risk (no remote downloads), but the bundle expects a Python virtualenv at CLAWWORK_DIR/venv and a larger ClawWork project at ~/.openclaw/workspace/ClawWork which the skill does not install. The wrapper sources the venv and runs local modules — missing setup steps could lead users to run unreviewed code from a different location.

⚠ Credentials

The code and SKILL.md require sensitive environment variables (OPENAI_API_KEY, OPENAI_API_BASE, and an E2B_API_KEY) but the skill's declared requirements list none and primary credential is unset. Requesting LLM and E2B keys is proportionate to the functionality, but the omission from metadata and the practice of loading a local .env file means secrets may be placed in a file under the user home and consumed by the skill without explicit declaration — a transparency and principle-of-least-privilege issue.

✓ Persistence & Privilege

The skill does not request always:true and does not modify other skills or system-wide agent settings. It writes a temporary config file (then deletes it) and reads/writes only under the ClawWork workspace path. No elevated persistence or cross-skill config changes are present in the bundle.

如何使用

确保已安装 OpenClaw（本地或 Docker 部署）
在对话框中输入安装命令：/install clawwork
安装完成后，直接呼叫该 Skill 的名称或使用 /clawwork 触发
根据 Skill 的参数说明提供必要输入，即可获得结构化输出

版本历史

v1.0.0

Initial release of ClawWork skill. - Integrates ClawWork to execute professional tasks with economic tracking, supporting over 220 tasks across 44 sectors. - Allows users to pay for tokens used and earn virtual "money" for completing high-quality tasks. - Provides CLI and direct usage options for task execution, model comparison, and agent status checks. - Includes setup instructions for API keys and dashboard access. - Offers detailed task economics and quality evaluation using models like GLM-4.7, Kimi K2.5, and Qwen3 Max.

元数据

Slug clawwork

版本 1.0.0

许可证 —

累计安装 0

当前安装数 0

历史版本数 1

常见问题

Clawwork 是什么？

Execute tarefas profissionais via ClawWork - transforme Zero em um coworker de IA economicamente viável. Use quando precisar executar trabalhos complexos, ge... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 557 次。

如何安装 Clawwork？

在 OpenClaw 或 Claude Code 对话框中运行命令「/install clawwork」即可一键安装，无需额外配置。

Clawwork 是免费的吗？

是的，Clawwork 完全免费（开源免费），可自由下载、安装和使用。

Clawwork 支持哪些平台？

Clawwork 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。

谁开发了 Clawwork？

由 felipetruman（@felipetruman）开发并维护，当前版本 v1.0.0。