ai-workflow-engine
ReviewAudited by ClawScan on May 10, 2026.
Overview
The skill mostly matches an AI workflow engine, but it tries to load undeclared code from hard-coded local OpenClaw skill paths, so it should be reviewed before use.
Review this skill before installing or running it. The broad automation features are expected for an AI workflow engine, but the hard-coded imports from local OpenClaw skill directories should be fixed or clearly declared. If you use it, run generated workflows in a sandbox, review code before execution, and protect API keys, database passwords, documents, and webhook payloads.
Findings (5)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Installing or using the skill could cause it to rely on, or execute, other local skill code that was not included in this review.
The module imports code from hard-coded local OpenClaw skill directories outside this skill's manifest. Python imports execute module top-level code, so if those paths exist, undeclared code may run when ai_workflow is imported.
sys.path.insert(0, r'C:\Users\qiuwe\.openclaw\.openclaw\workspace\skills\data-parser') from data_parser import DataParser as _DP
Ask the maintainer to remove hard-coded absolute paths, declare reviewed dependencies explicitly, and use package-relative imports or user-approved integrations.
A generated workflow could perform real actions such as writing to databases, sending messages, or publishing content if the user runs it.
The skill is designed to generate and execute workflows from natural-language descriptions, including examples for scraping, database operations, Excel output, email, and publishing.
根据你的简单描述,自动生成完整工作流!... # 执行 workflow.run()
Review generated workflow code and require explicit confirmation before running steps that write data, send email, publish content, scrape sites, or call external services.
If secrets are hard-coded or shared with generated workflows, they could be exposed through logs, files, or unintended service calls.
The skill supports provider API keys, database credentials, and SMTP configuration, which are expected for the stated integrations but are sensitive.
Config.set("openai_key", "sk-xxx")
Config.set("db_url", "mysql://user:pass@localhost/db")
Config.set("smtp", {"host": "smtp.gmail.com", "port": 587})Use least-privilege credentials, prefer a secret manager or environment variables, and avoid embedding real passwords or API keys in workflow code.
Documents or prior conversation content may influence later answers or workflows if added to the knowledge base or memory system.
The skill advertises memory and RAG features that can store or reuse user documents and conversation history.
记忆系统 - 短期/长期记忆 ... RAG知识库 - 文档向量化 - 多轮对话 - 记住对话历史
Only add intended documents, separate sensitive knowledge bases, and verify how memory is cleared or isolated between tasks.
Workflow data could be sent to external systems if a workflow calls a webhook with sensitive content.
The skill includes a generic webhook caller that can send arbitrary JSON data and headers to a user-supplied URL.
def call_webhook(self, url, method="POST", data=None, headers=None):
import requests
return requests.request(method, url, json=data, headers=headers or {})Confirm webhook destinations and payloads before use, and do not include secrets or private data unless the endpoint is trusted.