Skill flagged — suspicious patterns detected
ClawHub Security flagged this skill as suspicious. Review the scan results before using.
Clawtrade Bnb
v1.1.0Autonomous DeFi trading agent for BNB Chain with multi-strategy engine, network switching, and reinforced learning.
⭐ 0· 615·0 current·0 all-time
byAlan Estrada@alannetwork
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Suspicious
high confidencePurpose & Capability
Name/description (autonomous DeFi trading on BNB) match the shipped code: scheduler, tx executor, network switcher, reinforced learning, dashboard and on-chain logging are present. However the registry metadata advertises no required environment variables or primary credential while the code+docs clearly expect an RPC URL, a wallet private key (PRIVATE_KEY), and notification tokens. That discrepancy is disproportionate and inconsistent.
Instruction Scope
SKILL.md explicitly instructs the operator to provide a PRIVATE_KEY via env or .env and to edit RPC endpoints and contract addresses. The runtime flow includes reading/writing config and JSONL logs, emitting events, and executing real blockchain transactions (tx-executor.js). The docs also promote deploying a public dashboard/API (api/logs.js) that reads execution-log.jsonl—if logs or configs include sensitive fields they could be exposed. Instructions give the agent the ability to switch to mainnet and submit signed transactions; this is within the stated purpose but high-impact and the instructions for secret handling are risky and not reflected in declared requirements.
Install Mechanism
No external install spec or remote download is used (instruction-only installer). All code is bundled in the skill. That means nothing is pulled from arbitrary URLs at install time, which lowers supply-chain-install-time risk. However the package contains production-grade transaction-execution code (no external download needed) so installing it grants functionality to sign and send transactions locally.
Credentials
The skill metadata lists no required env vars, but SKILL.md and multiple files expect PRIVATE_KEY, RPC endpoints, and notification credentials (e.g., Telegram bot token / chat ID inside config.scheduler.json and notifications.js). Required secrets for executing transactions and sending alerts are therefore undeclared in the registry manifest; that mismatch is a security and transparency problem. The skill will request sensitive credentials (private key) that should have been declared as primaryEnv and documented as required—but they are not.
Persistence & Privilege
always:false (no forced installation) and model invocation is allowed (platform default). The skill does not request system-wide privileges in metadata. Nevertheless, it is autonomous-capable and contains code to sign/submit transactions and to run periodic scheduler cycles; combined with undeclared secrets this increases blast radius if run unintentionally or with a mainnet private key.
What to consider before installing
This skill is functional and can sign and submit real BNB-chain transactions, but the registry metadata does not declare the sensitive environment variables it needs (for example PRIVATE_KEY, RPC URL, and notification tokens). Before installing or running it:
- Do not run on mainnet until you have fully audited the code and removed any hardcoded secrets. Start on testnet only.
- Treat PRIVATE_KEY as highly sensitive: prefer a hardware wallet or a multisig/remote signer rather than storing raw keys in .env or env vars.
- Inspect tx-executor.js, scheduler.js, and notifications.js to confirm what is logged and what is sent to external services; ensure logs (execution-log.jsonl) do not leak secrets before exposing them via the dashboard/API.
- Add or require explicit env declarations (RPC_URL, PRIVATE_KEY, TELEGRAM_TOKEN/CHAT_ID, etc.) in the skill manifest so the platform can warn users and gate access.
- Consider running the skill inside an isolated environment with limited funds and monitoring, and add human-in-the-loop confirmations for mainnet actions (or enforce multi-sig) before trusting it with meaningful capital.
If you want, I can list the specific files that handle signing, logging, and notification (tx-executor.js, scheduler.js, agent-cli.js, on-chain-logger.js, notifications.js, api/logs.js) and point to lines or patterns you should review first.Like a lobster shell, security has layers — review code before you run it.
latestvk97b6jgm6y37t5cttphkjtcwbh81epr9
License
MIT-0
Free to use, modify, and redistribute. No attribution required.