Back to skill
Skillv0.1.0

ClawScan security

Mx Macro Data · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 4, 2026, 1:24 PM
Verdict
Benign
Confidence
high
Model
gpt-5-mini
Summary
The skill's code, instructions, and required EM_API_KEY align with its stated purpose (querying 东方财富 / EastMoney macro data); requested resources and network calls are proportional and there are no obvious exfiltration endpoints or excessive privileges.
Guidance
This skill appears coherent with its stated purpose, but take these precautions before installing: - Verify the EM_API_KEY source and scope: ensure the key is issued by the official EastMoney endpoint (https://ai.eastmoney.com/mxClaw), confirm its permissions and that you can revoke/regenerate it. - Run in an isolated environment (virtualenv, container) when first testing and install httpx there. - Inspect the included script (scripts/get_data.py) yourself — it posts queries to https://ai-saas.eastmoney.com/proxy/b/mcp/tool/searchMacroData and writes CSV/text files under a local miaoxiang/mx_macro_data folder. If that output location is sensitive, change it or run in a safe working directory. - Ensure network egress to ai-saas.eastmoney.com is acceptable in your environment and that TLS egress is allowed (the URL uses https). - Treat EM_API_KEY like any secret: do not embed it in code or logs, and revoke it if you believe it was compromised. - Note the small metadata inconsistency (registry says no install spec while SKILL.md includes a pip install instruction). Confirm dependency installation (httpx) before running. If you want, I can summarize the exact network calls and files the script will create or point out any lines in the code you should review closely.

Review Dimensions

Purpose & Capability
noteName/description claim natural-language macro data queries from 东方财富; the Python script posts to ai-saas.eastmoney.com/proxy/b/mcp/tool/searchMacroData using an EM_API_KEY header and writes CSV/description files. That matches the stated purpose. Minor inconsistency: registry metadata said 'No install spec / instruction-only' while the SKILL.md contains an openclaw 'install' entry for pip (httpx) and a code file (scripts/get_data.py) is included — operationally coherent but metadata bookkeeping is inconsistent.
Instruction Scope
okSKILL.md instructions and the script are limited to building a query, calling the EastMoney endpoint, converting returned JSON to CSV, and writing a description text file. The SKILL.md additionally mandates orchestrator-side verification (gap detection and iterative re-calls) which increases orchestration requirements but is within scope for a data-fetching skill. The skill does not instruct reading unrelated system files or other environment variables.
Install Mechanism
noteNo platform-level installer in the registry, but SKILL.md requests pip install httpx (a standard PyPI package). Installing httpx via pip is standard and low-risk. The mismatch between 'no install spec' in registry metadata and the SKILL.md install entry is a bookkeeping inconsistency to be aware of.
Credentials
okOnly a single environment variable (EM_API_KEY) is required and it is justified: the script places the key in an 'em_api_key' request header to authenticate to EastMoney. There are no other secret/env requests or unrelated credential reads.
Persistence & Privilege
okThe skill does not request always:true or modify other skills or system-wide config. It writes output files (CSV and description) to a local directory (miaoxiang/mx_macro_data by default) which is expected for a data exporter; this is a normal file-writing behavior and within the skill's scope.