Security audit
Security checks for vulnerabilities and agentic risk
This is a coherent takeout-coupon skill, but users should understand it contacts a third-party coupon API, opens commercial redemption links, and uses a persistent anonymous install identifier for attribution.
Install only if you are comfortable with the skill contacting agskills.moontai.top and sending a stable anonymous install ID for attribution. Treat coupon links and QR codes as third-party commercial destinations, verify the page before ordering, and consider deleting scripts/.install_id if you want to reset the local identifier.
The code automatically sends a persistent install identifier to a remote API as query parameters on every request without user-facing warning. This allows the operator to track a specific installation over time, linking usage patterns and potentially associating them with IP addresses, device context, and other request metadata.
The skill name, description, and all example invocations are presented exclusively in Chinese, implying the skill is intended to operate in a fixed language. The file does not offer users any language or locale choice, nor does it justify a China-specific locale restriction as required to avoid a language-policy issue.
npx commands without a version suffix (e.g. @1.0.0) create a rug-pull risk if the upstream server is compromised and publishes a malicious update.
The README explicitly states that API requests automatically append ch and u attribution parameters, including an anonymous installation identifier, for channel statistics. This creates undisclosed tracking behavior in a coupon-fetching skill and can enable user correlation, privacy leakage, and non-essential data collection beyond the skill's core function.
Suspicious Unicode normalization or mixed-script content
The skill explicitly states that requests include channel and installation UUID attribution parameters for anonymous install statistics and order attribution, but it does not clearly warn users before data is sent to the third-party API. This creates a transparency and privacy-consent issue because users may invoke the skill without understanding that tracking-related parameters are being transmitted.
The skill provides direct H5 redemption links and QR-code destinations that send users to third-party pages where they may redeem offers or place orders, but it does not give a clear safety warning about leaving the assistant context. This increases phishing, affiliate-abuse, and informed-consent risk because users may trust the links as first-party without understanding they are external commercial destinations.
Suspicious Unicode normalization or mixed-script content
The top-level description presents the script as a simple coupon fetcher, but the implementation also generates and persists a per-install UUID used for attribution tracking. This hidden tracking behavior undermines informed consent and creates a deceptive privacy gap between what users are told and what the code actually does.
The script's natural-language descriptions and output strings are entirely in Chinese, and the CLI/help text does not indicate any language choice or opt-in. For a general-purpose skill, this creates a locale/language constraint that is not user-selectable or justified in the file.
The script stores a persistent installation identifier in a local file without notifying the user. Even if not directly sensitive on its own, a stable identifier enables long-term correlation of activity across runs and can become privacy-invasive when combined with server-side logs or other metadata.
This markdown file is natural-language documentation, and it forces a single language/locale throughout the entire skill reference. Under the policy, a language constraint should either be user-selectable or clearly documented as region-specific and justified; neither is stated here.
