Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Antom Copilot
v2.0.0Antom Intelligent Assistant - Central control for all Antom-related requirements, intelligently analyzes user intent and delegates to the Payment Success Rat...
⭐ 18· 94·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
medium confidencePurpose & Capability
Name/description claim central Antom control and delegating to a Payment Success Rate Expert — the bundled Python scripts implement exactly that (pull data, analyze, generate PDF, send email). Requesting merchant_token and SMTP credentials via a local ~/antom/conf.json is consistent with the described features. However, the package also includes an embedded api_config.json with a bot token and an external API endpoint (ibotservice.alipayplus.com) that will be used by the scripts; shipping a hardcoded service token in the skill is unexpected and should be validated with the skill author.
Instruction Scope
Runtime instructions and scripts read/write files under ~/antom, parse conf.json containing merchant_token and SMTP username/password, and POST data containing the merchant token to an external API endpoint. That data exfiltration (merchant_token) is expected for the function, but it is sensitive — the skill transmits it to the included external endpoint. Also the scripts contain truncated/buggy code paths (e.g., undefined variable 'd' in JSON extraction, inconsistent implementations for data validity) which may cause unexpected failures or leaking of intermediate files. Review the code paths that extract and log API responses before supplying real credentials.
Install Mechanism
No installer; this is an instruction + code bundle. Dependencies are typical Python packages listed in requirements.txt (requests, matplotlib, reportlab, numpy). No remote downloads or install scripts are executed by the skill itself.
Credentials
The skill does not request environment variables but requires a local conf.json containing merchant_token and SMTP credentials (username/password). Those are proportionate to the stated capabilities (API auth and sending email) but are sensitive. Additionally, the skill includes its own api_config.json with a hardcoded bot token and API endpoint — an embedded credential that will be used by the skill but is not documented as a required secret in the SKILL metadata.
Persistence & Privilege
The skill does not request always:true and does not modify other skills. It will create and write files under ~/antom and ~/antom/success rate which is consistent with report generation and expected file outputs.
What to consider before installing
What to consider before installing/running:
- Code review: Inspect api_config.json and all Python scripts before running. The skill ships with a hardcoded bot token and an external API endpoint (ibotservice.alipayplus.com) — confirm you trust that service and that the token is intended to be public. Ask the author where that token comes from.
- Secrets handling: The scripts expect a local ~/antom/conf.json that contains merchant_token and SMTP username/password. Do NOT put production credentials there until you've audited the code. Prefer using an app-specific email password or a throwaway account for testing.
- Sandbox first: Run the scripts in an isolated environment (throwaway VM, container) and with test credentials to observe network calls and output files. Monitor outbound requests to ensure only the documented endpoint is contacted.
- Audit bugs and behavior: The code has implementation inconsistencies (different is_valid_data logic across scripts) and at least one apparent bug in JSON extraction (undefined variable). These indicate the package may be untested — expect runtime errors and verify error-handling so secrets aren't leaked to logs.
- Operational controls: If you decide to use it, rotate any credentials that were exposed to the skill, and configure network egress rules (or allowlist only the required endpoint) if possible.
If you want, I can list all locations in the code that read or transmit sensitive values (merchant_token, SMTP password) and the exact network endpoints they call so you can focus your review.Like a lobster shell, security has layers — review code before you run it.
latestvk97aavjj8zwd2ag4cg535t25qh83q7vv
License
MIT-0
Free to use, modify, and redistribute. No attribution required.
