Skillv1.0.0

ClawScan security

双币赢分析 · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

SuspiciousMar 5, 2026, 12:14 PM

Verdict: suspicious
Confidence: medium
Model: gpt-5-mini
Summary: The skill delegates core analysis to an undisclosed remote HTTP service (hard-coded IP) and the SKILL.md does not disclose or justify this network dependency — this mismatch is concerning and deserves caution.
Guidance: This skill calls a hard-coded HTTP endpoint (43.156.132.183:30080) to obtain all analysis results but does not disclose that in its description or SKILL.md. Because responses come from an external, unencrypted, untrusted server, the provider can return arbitrary output (including misleading or harmful trading recommendations) and can observe the agent's network requests. Before installing or using: (1) ask the publisher for the source and justification for using that specific API and for a TLS-secured domain; (2) request a self-contained/local implementation or an option to run analysis locally; (3) if you must run it, sandbox network access and do not use it for live trading decisions or with real credentials/funds; (4) verify presence of required Python dependencies (requests) and consider auditing the remote service or running in an isolated environment. If you cannot verify the remote service's trustworthiness, treat this skill as high-risk and avoid using it for real financial decisions.

Purpose & Capability: noteThe skill's stated purpose (BTC/ETH dual-win options analysis) matches the delivered behavior: it returns market/strategy analysis. However, instead of performing local calculations described in SKILL.md, the skill fetches results from a remote service, which is not documented in the description.
Instruction Scope: concernSKILL.md describes detailed local decision logic and polling behavior, but runtime behavior (run.py) is a single HTTP GET to an external API that could return any content. The documentation does not disclose that analysis is performed remotely, nor that network calls to an arbitrary IP will occur.
Install Mechanism: okNo install spec — instruction-only with a small code file. Nothing is written to disk beyond the provided files. Lower risk from install perspective.
Credentials: concernThe skill requests no environment variables or credentials, but it makes an outbound HTTP request to a hard-coded IP (http://43.156.132.183:30080). The outbound network dependency is not declared in the SKILL.md or metadata and uses plain HTTP (no TLS), which is disproportionate given the apparent local-analysis nature of the documentation.
Persistence & Privilege: okalways is false and the skill does not request persistent privileges or modify other skills. It runs as an executable action with a 60s timeout — no elevated platform privileges requested.