Back to skill
Skillv1.0.0

ClawScan security

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

Scanner verdict

BenignFeb 15, 2026, 3:33 PM
Verdict
benign
Confidence
high
Model
gpt-5-mini
Summary
The skill is internally consistent: it is a Node CLI for interacting with a self‑hosted Mealie API and only requests the Mealie URL and API token it needs.
Guidance
This skill appears to do exactly what it says: a Node CLI that calls your Mealie instance using MEALIE_URL and MEALIE_API_TOKEN. Before installing: (1) confirm you trust the skill’s source since it will store/read a token in ~/.openclaw/.env or the skill .env; (2) prefer creating a limited-purpose API token in Mealie rather than using an admin token; (3) ensure MEALIE_URL uses HTTPS on a trusted host to avoid token exposure; (4) review the included script yourself if you are concerned (it is small and readable); and (5) avoid putting other secrets in the agent-level .env since the skill will read that file for MEALIE_* variables.

Review Dimensions

Purpose & Capability
okThe name/description (Mealie recipe manager client) matches the declared requirements (node, MEALIE_URL, MEALIE_API_TOKEN) and the included script implements the described API calls (recipes, lists, meal plans).
Instruction Scope
okSKILL.md and the CLI commands are limited to calling the Mealie API. The script reads only MEALIE_URL and MEALIE_API_TOKEN (from environment or .env files) and makes HTTP(S) requests to the Mealie host; it does not reference other system paths or external endpoints.
Install Mechanism
okNo install spec or remote downloads are present; the skill is instruction + a local Node script. Nothing in the manifest pulls arbitrary code from external URLs.
Credentials
okThe skill requires only MEALIE_URL and MEALIE_API_TOKEN, which are precisely the credentials necessary for a Mealie client. The code explicitly loads only these two variables from skill-level and agent-level .env files.
Persistence & Privilege
okalways is false and the skill does not request elevated/system-wide privileges or modify other skills; it only reads its own /agent .env locations as documented.