Bsc Dev Monitor Skill
WarnAudited by ClawScan on May 10, 2026.
Overview
The skill’s BSC monitoring purpose is mostly coherent, but it ships hardcoded ClawHub and payment credentials and has inconsistent billing terms.
Do not run the deployment scripts, and do not rely on this version until the exposed credentials are removed and rotated. Confirm the exact billing price/mode before invoking it, and use only trusted webhook URLs with bounded monitor durations.
Findings (6)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Anyone with the package could potentially abuse the exposed ClawHub account, and running the deployment script would use credentials unrelated to normal wallet monitoring.
The artifact contains an actual hardcoded ClawHub login password, redacted here, and the script uses it to authenticate and upload a skill.
const CREDENTIALS = { email: 'hefang080@gmail.com', password: '[hardcoded password]' };Revoke the exposed password, remove account credentials from the package, and use user-specific login/OAuth or environment-managed secrets outside the distributed skill.
The payment key could be copied from the skill and misused for billing-related API calls or leaked through URLs/logs.
A SkillPay API key is embedded in runtime code and is also placed into generated payment-link query data, exposing payment-provider authority to anyone who can inspect or receive the package/output.
apiKey: 'sk_f072…b994b5' ... paymentData = { api_key: SKILLPAY_CONFIG.apiKey, ... }Rotate the SkillPay key, keep billing secrets server-side, and do not include secret API keys in SKILL.md, source code, or payment URLs.
Users may agree to the skill expecting one price or billing mode and encounter different billing behavior.
The runtime code references 0.01 USDT and a per-detection default, while SKILL.md advertises 0.001 USDT per call, making the actual charge behavior ambiguous.
price: '0.01', ... billingMode: 'per_call' ... billing_mode = 'per_detection'
Publish one clear billing model and price across SKILL.md, code, metadata, and deployment docs before users install or invoke the skill.
Running deployment helpers could change the local environment or publish/upload code unintentionally.
The deployment helper can install a global package and publish the current directory if a user or agent runs it, though no install spec indicates it runs automatically.
npm install -g @clawhub/cli ... clawhub publish .
Do not run deployment scripts as part of normal skill use; keep deployment tooling separate, pinned, and credential-free.
A monitor may keep polling and sending notifications until its duration expires or it is stopped.
Creating a monitor starts a repeated background loop. This is purpose-aligned and SKILL.md documents duration/stop controls, but users should notice that monitoring can continue over time.
monitorLoop(monitorId); ... setTimeout(() => monitorLoop(monitorId), 10000);
Use bounded durations when possible and explicitly stop monitors you no longer need.
The configured webhook endpoint can learn which addresses/tokens you monitor and when detections occur.
The skill supports automatic webhook notifications to a user-supplied URL, which is expected for alerts but sends monitored-event data outside the skill.
"webhook_url": "https://your-server.com/webhook" ... 系统会自动推送
Use only trusted HTTPS webhook endpoints and consider adding webhook signing or authentication.