Skillv1.0.0

ClawScan security

cross-border-finance · ClawHub's context-aware review of the artifact, metadata, and declared behavior.

Scanner verdict

BenignApr 20, 2026, 7:54 AM

Verdict: Benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's requested artifacts and runtime instructions are consistent with a cross‑border cost‑calculator and recommendation assistant; nothing in the package demands unrelated secrets or installs arbitrary code from external hosts.
Guidance: This skill appears to do what it says: compute fees and recommend paths using a bundled Python script and static reference docs. Before installing, note two practical risks: (1) the skill explicitly describes and may recommend regulatory 'gray' options (P2P crypto, off‑ramp workarounds) — these carry real legal and bank‑risk consequences; do not treat recommendations as legal or compliance advice and consider consulting a lawyer/tax advisor for large amounts; (2) the calculator fetches live FX rates from a public API (frankfurter.app) when run, so it requires outbound network access — if you run the script in a restricted environment be aware of that call. If you want stronger assurance, review the full calculate_costs.py contents locally before running and validate that it makes only the expected HTTP requests (exchange rate API) and does not transmit user data to any other endpoints.

Review Dimensions

Purpose & Capability: okName/description (cross‑border fees and path recommendations) match the included assets: an analysis script (calculate_costs.py) and multiple reference docs about banks and crypto platforms. No unrelated binaries, credentials, or config paths are requested.
Instruction Scope: noteSKILL.md instructs the agent to collect user inputs and run the included Python script which fetches live FX rates (frankfurter.app) and computes comparisons. That is within scope. Note: the skill explicitly includes and recommends '灰色/ P2P 加密' paths and steps for potentially risky workflows (P2P crypto, splitting flows to avoid on‑chain checks). Those suggestions align with the stated purpose but have legal/regulatory implications that the skill already flags. The instructions do not ask the agent to read unrelated files or environment variables.
Install Mechanism: okNo install spec; this is instruction + included script and static references. There is no download URL or extract step. The only network call is to a public FX API (frankfurter.app) invoked by the script, which is expected for live rate lookups.
Credentials: okThe skill declares no required env vars, no credentials, and no config paths. The included code does not require secrets to run. No disproportionate credential requests are present.
Persistence & Privilege: okalways:false and the skill does not request system‑wide privileges or to modify other skills. It is user‑invocable and may be autonomously invoked by the model (platform default), which is appropriate for this type of tool.