Back to skill
Skillv1.0.0

ClawScan security

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

Scanner verdict

ReviewMar 15, 2026, 2:08 PM
Verdict
Review
Confidence
medium
Model
gpt-5-mini
Summary
The skill mostly matches its wine-management purpose, but there are inconsistencies between the documentation and the actual files (missing advertised scripts/files and truncated code), so inspect and test before trusting it with your data.
Guidance
This skill is not obviously malicious, but there are inconsistencies and missing pieces you should consider before installing: - The documentation claims several helper scripts and data files (e.g., add_wine.py, log_consumption.py, generate_report.py, consumption_log.json, purchase_history.json) that are not present. That may mean the skill is incomplete or the docs are out of date. Verify the exact behavior you need (e.g., logging consumption) by testing the included scripts first. - The included scripts operate only on JSON files inside the skill directory (data/). They will create or overwrite files there via save_json. Back up any important data and run the skill in an isolated environment if possible. - There are no network calls in the shipped code, but lookup_barcode.py contains TODO comments to call external wine APIs in future. If the skill is updated to call external services it may require API keys; treat such updates as security-relevant and review any new network code and required environment variables before enabling. - Recommended immediate steps: review and run the included scripts in a sandbox, confirm they only touch the skill's data/ folder, and verify that recommended features you expect are implemented. If you need the missing advertised scripts, ask the publisher for a complete release or inspect updates closely. If you will allow autonomous invocation in an agent, be aware the skill can read/write the local skill data files—ensure that is acceptable for your environment.

Review Dimensions

Purpose & Capability
noteThe code present (barcode lookup, recommendation engine, utils) supports core features described (inventory, barcode lookup, meal pairing). However SKILL.md advertises additional scripts and artifacts (add_wine.py, log_consumption.py, generate_report.py, data/consumption_log.json, data/purchase_history.json, many reference docs) that are not present in the file manifest. Some functionality is implemented under different names (e.g., add_wine_from_barcode in lookup_barcode.py) but the mismatch between advertised and included files is notable.
Instruction Scope
okRuntime instructions and scripts operate only on JSON files in the skill's data directory and local reference data. The SKILL.md does not instruct reading unrelated system files or contacting remote endpoints. The scripts likewise do local file I/O and in-code lookups; lookup_barcode references external APIs only as TODOs and does not perform network calls.
Install Mechanism
okNo install specification is present (instruction-only skill with included scripts). Nothing is downloaded or installed at runtime according to the manifest—this reduces install-time risk.
Credentials
okThe skill requests no environment variables, no credentials, and no config paths. The code does not attempt to access external credentials. This is proportional for a local inventory/recommendation tool.
Persistence & Privilege
okThe skill is not forced-always, and it does not modify other skills or system-wide settings. It persists data only under its own data/ directory which is consistent with its stated storage behavior.