Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
IELTS Reading Review 雅思阅读复盘助手
v1.2.0IELTS Reading passage review, scoring, and progress tracking skill. This skill should be used when the user finishes an IELTS Academic Reading passage or ful...
⭐ 0· 103·0 current·0 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
Name/description match the files: HTML template, PDF generator, and review references support the stated review, scoring, and tracking features. However, the included Cloudflare worker and hard-coded telemetry endpoint are not required for core functionality (review, HTML/PDF output, band conversion) and therefore appear optional extra components rather than necessary capabilities.
Instruction Scope
SKILL.md stays on scope for collecting passage text, answers, and producing HTML/PDF output and progress tables. It also promises 'cross-session' accumulation of vocab/error patterns, but does not clearly explain how persistence is implemented (no declared backend or credentials). This gap is ambiguous: persistence could be via the agent platform's memory, but the README/code point to an external stats worker only for anonymous telemetry (not per-user data).
Install Mechanism
There is no automatic install spec (instruction-only skill), which reduces installation risk. Files include a Node.js PDF script (puppeteer-core) and optional Cloudflare deployment scripts; these are reasonable for HTML→PDF and optional telemetry. Nothing here downloads arbitrary binaries from unknown hosts, but running the scripts requires installing Node/puppeteer and (for the worker) Cloudflare Wrangler.
Credentials
The skill declares no required environment variables, which is good. Nonetheless, the PDF script contacts a hard-coded third-party endpoint (https://ielts-skill-stats.dengjiawei.workers.dev/ping) to report anonymous events by default. That telemetry is not necessary for core functionality and creates a privacy surface. The README/script allow disabling telemetry via SKILL_NO_TELEMETRY=1, but the default behavior is to call the external endpoint. The stats-worker deploy script requires Cloudflare credentials if you choose to deploy it — appropriate for an optional telemetry backend but unrelated to core review functionality.
Persistence & Privilege
The skill claims cross-session accumulation of mistakes and vocabulary. There is no code here that securely implements per-user persistent storage under the user's control: the included worker is for aggregate telemetry (author-controlled domain) and does not implement per-user vocab/error storage. If the skill relies on the agent's memory for persistence that is platform-dependent and should be made explicit. Also, telemetry to the author's endpoint will reveal network-level metadata (IP, request timing) to that endpoint even if payload claims to be 'anonymous.'
Scan Findings in Context
[external-telemetry-endpoint] unexpected: generate-pdf.js POSTs anonymous events to https://ielts-skill-stats.dengjiawei.workers.dev/ping. Telemetry is plausible for analytics, but it is not required for review or PDF generation and is enabled by default (can be disabled via SKILL_NO_TELEMETRY=1).
[puppeteer-core-usage] expected: scripts/generate-pdf.js uses puppeteer-core and local Chrome to convert HTML to PDF — expected for an optional PDF export feature. Running the script requires Node and a local Chrome binary.
[cloudflare-wrangler-deploy] expected: stats-worker/deploy.sh and worker.js are provided to deploy an optional Cloudflare Worker (KV-backed) for collecting statistics. This is an optional telemetry backend and requires the user's Cloudflare account/credentials to deploy.
[claims-anonymous-but-network-exposes-ip] unexpected: Worker and README state 'anonymous' usage data only (version + timestamp). In practice any HTTP endpoint will see the client's IP and other network metadata; this is a privacy caveat not acknowledged in the code comments.
What to consider before installing
This skill appears to implement the stated review, HTML output, and optional PDF export features. Before installing or running it:
- If you will use the PDF script (scripts/generate-pdf.js), be aware it will attempt to call an external telemetry endpoint by default. To avoid outgoing telemetry calls set SKILL_NO_TELEMETRY=1 in your environment before running the script. Review the script source before running and install puppeteer/core and a local Chrome only if you trust it.
- The included Cloudflare worker and deploy script are optional telemetry infrastructure. You do NOT need to deploy that worker to use the core skill. If you do deploy it, you will be granting Cloudflare Wrangler access to your Cloudflare account and hosting a third-party endpoint that will receive usage events — review the worker code and wrangler configuration carefully.
- The skill promises cross-session accumulation of vocabulary and errors but does not include a clear, private per-user storage mechanism; ask the skill author or the platform how memory/persistence is implemented and where user-specific review data will be stored. Do not upload personally identifiable information or sensitive exam/test copies if you are unsure where they will be stored or transmitted.
- If you are comfortable with the code, you can keep telemetry disabled (SKILL_NO_TELEMETRY=1) and generate PDFs locally. If you are not comfortable, run the skill in a sandbox or inspect/modify the generate-pdf.js to remove the network call before use.
Overall: functional and mostly coherent, but telemetry and unclear persistence warrant caution — review and disable telemetry if you prefer not to send usage events to the author's endpoint.Like a lobster shell, security has layers — review code before you run it.
latestvk975nstfaym5vdhwja67j6adjx841tmf
License
MIT-0
Free to use, modify, and redistribute. No attribution required.