Skill flagged — suspicious patterns detected

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

Customer Logo Wall

v1.1.0

处理客户表格并生成Logo墙PPT的自动化工具。当用户提到"客户表格"、"公司中文名"、"Logo墙"、"客户logo"或需要处理Excel中的客户数据并生成PPT时使用此技能。

0· 73·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
VirusTotalVirusTotal
Suspicious
View report →
OpenClawOpenClaw
Suspicious
medium confidence
!
Purpose & Capability
The declared registry metadata lists no required binaries or env vars, but the SKILL.md and scripts clearly require Node/Playwright, Chrome, Python packages, and optionally tesseract. Worse: SKILL.md insists the tool must read the user's Excel and write the Chinese-name column back into the original Excel file, but none of the provided scripts implement reading/writing Excel (there is no script to convert Excel→JSON or to update the original XLSX). That is a major functional mismatch between claimed behavior and included code.
!
Instruction Scope
Runtime instructions explicitly require modifying the user's original Excel file ("必须 将结果写回用户的原始 Excel 文件") which is high-impact and surprising for many users. The verification step uses subprocess calls to an AGENT_BROWSER executable to open pages and eval JS—this delegates DOM scraping to an external program whose identity and behavior are uncontrolled. The skill downloads arbitrary remote images and writes files to disk; these network and file-write behaviors are within the skill purpose but the Excel overwrite and external executable invocation expand the scope significantly.
Install Mechanism
There is no install spec (instruction-only + scripts), so nothing is automatically downloaded at install time — lower platform install risk. However, dependencies (Node/playwright, Chrome, Python libs, optional tesseract) are required but not declared in registry metadata, which is a packaging/metadata omission that can cause surprising manual setup steps for users.
!
Credentials
The skill requests no cloud credentials (which is appropriate) but expects environment variables (AGENT_BROWSER, NODE_PATH, NODE_BIN) to be set by the user. verify_logos.py disables SSL verification for urllib requests (SSL_CTX.verify_mode = ssl.CERT_NONE), reducing transport security when redownloading logos. The verify/redownload flow invokes an external AGENT_BROWSER executable via subprocess with 'open' and 'eval' commands — allowing any binary named in AGENT_BROWSER to execute arbitrary operations with the agent's privileges. Those behaviors are disproportionate unless the user deliberately configures a trusted browser-automation executable.
!
Persistence & Privilege
always:false and no platform-level persistence are fine, but the workflow mandates overwriting the user's original Excel file (writing user data back in-place). That is a destructive action relative to normal read-only data-processing expectations and should require explicit, visible user consent and backups before running.
What to consider before installing
Key things to consider before installing or running this skill: - Source trust: the skill's source/homepage is unknown. Only proceed if you trust the author or have reviewed the code. - Excel overwrite: SKILL.md says it will write the Chinese-name column back into your original Excel file and overwrite it. There is no provided script that reads/writes Excel in the bundle, so verify how Excel→JSON conversion and the actual overwrite would occur. Always back up your Excel files before running. - External executable (AGENT_BROWSER): verify what you set AGENT_BROWSER to. The verify script runs that executable with 'open' and 'eval' via subprocess — if that binary is malicious or compromised it can run arbitrary code. Prefer running in a sandbox/container or replacing that flow with a trusted scraper you control. - SSL verification: verify_logos.py disables SSL certificate checks for downloads. This weakens security and may let man-in-the-middle data be returned. Consider editing the code to use default SSL verification or run it on a network you trust. - Dependency & metadata mismatch: metadata in the registry claims no required binaries/env vars, but the README and scripts require Node, Playwright, Chrome, Python 3.9+, python-pptx, pillow, pytesseract (optional). Make sure dependencies come from trusted package sources. - Test safely: run the scripts on a small, disposable sample dataset in an isolated environment (container or VM), and inspect the outputs and any files written. If you need the Excel integration, implement or verify the Excel read/write step yourself and remove any automatic in-place overwrite unless explicitly desired. - Code fixes: redownload logic in verify_logos.py contains suspect/buggy code paths (e.g., string .includes used like JS), indicating the scripts haven't been robustly tested. Treat outputs cautiously. If you can't perform these checks, treat the skill as risky and avoid using it with sensitive customer data.

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

automationvk977zb54rw9pyn8px826z95sfn83jat8excelvk977zb54rw9pyn8px826z95sfn83jat8latestvk9766qmfs4ks290x12sab8pf5s843284logovk977zb54rw9pyn8px826z95sfn83jat8pptvk977zb54rw9pyn8px826z95sfn83jat8

License

MIT-0
Free to use, modify, and redistribute. No attribution required.

Comments