Skillv1.3.3

ClawScan security

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

Scanner verdict

BenignFeb 13, 2026, 9:22 AM

Verdict: benign
Confidence: high
Model: gpt-5-mini
Summary: The skill's requirements, files, and runtime instructions are consistent with a Loxone Miniserver client for querying and controlling smart-home devices; nothing in the bundle appears aimed at unrelated or covert data access, though there are a few minor documentation inconsistencies and normal security considerations around storing credentials.
Guidance: This package appears to do what it says: communicate with a Loxone Miniserver via HTTP and WebSocket. Before installing or running: 1) Note the mismatch in docs—SETUP.md incorrectly claims only standard library; you must install requests and websockets (from PyPI) or otherwise ensure they exist. 2) The skill stores your Loxone username/password in a local config.json—follow the SETUP.md advice and restrict file permissions (chmod 600). 3) The skill will contact dns.loxonecloud.com if you use Cloud DNS and will talk directly to the configured Miniserver; only provide credentials/remote access you trust. 4) Control commands can have real-world effects—use the read-only defaults and only run control actions when you explicitly intend to (the scripts include a safe-rooms check). 5) Consider running the scripts in an isolated environment (virtualenv/container) and review the code yourself if you have sensitive network exposure requirements. If you want higher assurance, ask the publisher to fix the documentation about required Python packages and confirm the expected config file path.

Review Dimensions

Purpose & Capability: okName/description match the included Python scripts (HTTP API client + WebSocket watcher) and the listed commands. The network calls (Miniserver endpoints and Loxone Cloud DNS) are expected for this purpose. Minor documentation mismatch: SETUP.md claims "No additional Python packages required (standard library)" while SKILL.md and the code require third-party packages (requests, websockets).
Instruction Scope: okSKILL.md and the scripts limit actions to Loxone-related activities: loading a local config.json, fetching LoxAPP3.json, opening HTTPS/HTTP to the configured host, and connecting to dns.loxonecloud.com when using Cloud DNS. The skill reads only its own config file and caches a structure file; it does not attempt to read unrelated system files or environment variables. It prints partial tokens to stdout during auth flows (informational).
Install Mechanism: okNo install spec (instruction-only at registry level) — the runtime is the included Python scripts. This is lower-risk than arbitrary remote installs. The scripts do depend on third-party Python packages (requests, websockets) — ensure these are installed from known sources before running.
Credentials: okNo environment variables or external credentials are requested by the registry metadata. The skill requires Loxone username/password, but these are stored in a local config.json as documented (expected for a network client). This credential usage is proportional to the stated functionality; users should protect the local config file (SETUP.md recommends chmod 600).
Persistence & Privilege: okalways:false and no claims of modifying other skills or system-wide settings. The skill caches a structure file in its directory and a small .cache LoxAPP3.json — normal for offline efficiency. Agent autonomous invocation is enabled by default but not elevated by the skill.