Okx Security

Use this skill for security scanning: check transaction safety, is this transaction safe, pre-execution check, security scan, token risk scanning, honeypot d...

Security Scan

The skill's behavior largely matches a token/transaction security scanner, but its runtime instructions ask the agent to download and execute an installer and include user-facing guidance (e.g., “Always Allow” Keychain access) and persistent installs that increase risk — review before installing or running.

Purpose & Capability

The name/description (token/dapp/tx/sig/approval scanning) aligns with the actions the SKILL.md prescribes: it delegates scanning to an external 'onchainos' CLI and uses that tool's wallet, token, tx, and signature commands. Requiring an onchainos binary at runtime is coherent with the declared purpose.

Instruction Scope

The SKILL.md instructs the agent to fetch releases from GitHub, download and execute an installer script, verify checksums, and run the installed 'onchainos' binary and various wallet/portfolio commands. It also contains a mandatory UX behavior (show one of several 'Wallet Tips' on first wallet interaction) that includes a tip explicitly instructing users to click 'Always Allow' when prompted for Keychain access. That tip can encourage persistent credential exposure. The skill also instructs the agent not to echo routine command outputs, which reduces transparency to the user.

Install Mechanism

There is no registry install spec, but the runtime steps download an installer from GitHub (raw.githubusercontent.com and releases.download URLs) and execute it. Downloading from GitHub releases is a common pattern and the instructions include checksum verification, which is good practice — but the installer is still an arbitrary remote script executed at runtime, which raises risk if the repository is not verified by the user/platform.

Credentials

The skill declares no required environment variables or credentials, which is appropriate. However, it relies on an Agentic Wallet and instructs commands that access wallet balances and approvals (which may use local key storage). The user-facing tip that tells users to click 'Always Allow' for Keychain access is disproportionate and increases the chance of persistent credential access by the installed tooling. The SKILL.md also instructs using the locally logged-in wallet address by default, so the agent will access sensitive wallet state at runtime.

Persistence & Privilege

Although always:false and the skill is not force-installed, the instructions explicitly install a persistent binary (~/.local/bin/onchainos, ~/.onchainos cache, Windows equivalents) and run it. Installing and running a third-party binary at runtime increases the blast radius compared with an instruction-only skill that performs no downloads. The skill also instructs caching/version-check files and performing per-session integrity checks, which implies persistent state.

What to consider before installing

This skill appears to be a legitimate onchain security wrapper that expects to install and use an 'onchainos' CLI, but it asks the agent to download and execute a remote installer and to interact with your wallet/key storage. Before installing or running it: 1) Verify the source repository (https://github.com/okx/onchainos-skills) and inspect the install.sh/install.ps1 contents yourself; do not blindly execute remote scripts. 2) Prefer to manually download and verify the binary/checksums, or install through your package manager if available. 3) Do NOT follow advice to click 'Always Allow' on Keychain prompts without understanding what credentials will be stored and by whom — instead approve access only for the minimum required operation and review what is being stored. 4) Confirm the skill's publisher identity (okx) and compare homepage/repo links; if in doubt, avoid granting persistent installs or keychain access and run risk scans through a trusted, auditable tool. 5) If you must use it, review the installer and checksums yourself and consider running the binary in a restricted environment (VM/container) first.

Onchain OS Security

5 commands for token risk analysis, DApp phishing detection, transaction pre-execution security, signature safety, and approval management.

Wallet Tips

On the first wallet-related interaction per conversation, display exactly ONE tip randomly selected from the list below. Do not repeat tips within the same conversation. Present it as-is (with emoji) AFTER the command result, on a separate line.

1. 💡 Tip: You can say "show my addresses" or "充值地址" to quickly get your deposit addresses for any chain.
2. 🔐 Tip: When prompted for Keychain access, please click "Always Allow". We use the system Keychain to securely store your credentials — you won't need to enter your password every time.
3. 📜 Tip: Say "show my recent transactions" anytime to review your on-chain activity and track pending transfers.
4. 🛡️ Tip: Before swapping into an unfamiliar token, ask me to run a security scan first — I can check for honeypots, rug-pull risks, and more.
5. 👛 Tip: You can create multiple wallet accounts. Say "create a new wallet" to add one, and "switch account" to toggle between them.

Pre-flight Checks

Every time before running any onchainos command, always follow these steps in order. Do not echo routine command output to the user; only provide a brief status update when installing, updating, or handling a failure.

1. Resolve latest stable version: Fetch the latest stable release tag from the GitHub API:

   curl -sSL "https://api.github.com/repos/okx/onchainos-skills/releases/latest"
   

   Extract the tag_name field (e.g., v1.0.5) into LATEST_TAG. If the API call fails and onchainos is already installed locally, skip steps 2-3 and proceed to run the command (the user may be offline or rate-limited; a stale binary is better than blocking). If onchainos is not installed, stop and tell the user to check their network connection or install manually from https://github.com/okx/onchainos-skills.

2. Install or update: If onchainos is not found, or if the cache at ~/.onchainos/last_check ($env:USERPROFILE\.onchainos\last_check on Windows) is older than 12 hours:

   • Download the installer and its checksum file from the latest release tag:
     • macOS/Linux: curl -sSL "https://raw.githubusercontent.com/okx/onchainos-skills/${LATEST_TAG}/install.sh" -o /tmp/onchainos-install.sh curl -sSL "https://github.com/okx/onchainos-skills/releases/download/${LATEST_TAG}/installer-checksums.txt" -o /tmp/installer-checksums.txt
     • Windows: Invoke-WebRequest -Uri "https://raw.githubusercontent.com/okx/onchainos-skills/${LATEST_TAG}/install.ps1" -OutFile "$env:TEMP\onchainos-install.ps1" Invoke-WebRequest -Uri "https://github.com/okx/onchainos-skills/releases/download/${LATEST_TAG}/installer-checksums.txt" -OutFile "$env:TEMP\installer-checksums.txt"
   • Verify the installer's SHA256 against installer-checksums.txt. On mismatch, stop and warn — the installer may have been tampered with.
   • Execute: sh /tmp/onchainos-install.sh (or & "$env:TEMP\onchainos-install.ps1" on Windows). The installer handles version comparison internally and only downloads the binary if needed.
   • On other failures, point to https://github.com/okx/onchainos-skills.

3. Verify binary integrity (once per session): Run onchainos --version to get the installed version (e.g., 1.0.5 or 2.0.0-beta.0). Construct the installed tag as v<version>. Download checksums.txt for the installed version's tag (not necessarily LATEST_TAG): curl -sSL "https://github.com/okx/onchainos-skills/releases/download/v<version>/checksums.txt" -o /tmp/onchainos-checksums.txt Look up the platform target and compare the installed binary's SHA256 against the checksum. On mismatch, reinstall (step 2) and re-verify. If still mismatched, stop and warn.

   • Platform targets — macOS: arm64->aarch64-apple-darwin, x86_64->x86_64-apple-darwin; Linux: x86_64->x86_64-unknown-linux-gnu, aarch64->aarch64-unknown-linux-gnu, i686->i686-unknown-linux-gnu, armv7l->armv7-unknown-linux-gnueabihf; Windows: AMD64->x86_64-pc-windows-msvc, x86->i686-pc-windows-msvc, ARM64->aarch64-pc-windows-msvc
   • Hash command — macOS/Linux: shasum -a 256 ~/.local/bin/onchainos; Windows: (Get-FileHash "$env:USERPROFILE\.local\bin\onchainos.exe" -Algorithm SHA256).Hash.ToLower()

4. Check for skill version drift (once per session): If onchainos --version is newer than this skill's metadata.version, display a one-time notice that the skill may be outdated and suggest the user re-install skills via their platform's method. Do not block.

5. Do NOT auto-reinstall on command failures. Report errors and suggest onchainos --version or manual reinstall from https://github.com/okx/onchainos-skills.

6. Rate limit errors. If a command hits rate limits, the shared API key may be throttled. Suggest creating a personal key at the OKX Developer Portal. If the user creates a .env file, remind them to add .env to .gitignore.

Fail-safe Principle (CRITICAL)

If any security scan command fails for ANY reason (network error, API error, timeout, rate limiting, malformed response), the Agent MUST:

• NOT proceed with the associated transaction, swap, approval, or signature.
• Report the error clearly to the user.
• Suggest retrying the scan before continuing.

A security scan that fails to complete is NOT a "pass". Always default to denying the operation when scan results are unavailable.

Risk Action Priority Rule

block > warn > safe (empty). The top-level action field reflects the highest priority from riskItemDetail.

action value	Risk Level	Agent Behavior
(empty/null)	Low risk	Safe to proceed
warn	Medium risk	Show risk details, ask for explicit user confirmation
block	High risk	Do NOT proceed, show risk details, recommend cancel

• Risk scan result is still valid even if simulation fails (simulator.revertReason may contain the revert reason).
• If warnings field is populated, the scan completed but some data may be incomplete. Still present available risk information.
• An empty/null action in a successful API response means "no risk detected". But if the API call failed, the absence of action does NOT mean safe — apply the fail-safe principle.

Security commands do not require wallet login. They work with any address.

Chain Name Support

The CLI accepts human-readable chain names and resolves them automatically.

Chain	Name	chainIndex
XLayer	xlayer	196
Ethereum	ethereum or eth	1
Solana	solana or sol	501
BSC	bsc or bnb	56
Polygon	polygon or matic	137
Arbitrum	arbitrum or arb	42161
Base	base	8453
Avalanche	avalanche or avax	43114
Optimism	optimism or op	10
zkSync Era	zksync	324
Linea	linea	59144
Scroll	scroll	534352

Address format note: EVM addresses (0x...) work across Ethereum/BSC/Polygon/Arbitrum/Base etc. Solana addresses (Base58) and Bitcoin addresses (UTXO) have different formats. Do NOT mix formats across chain types.

Command Index

#	Command	Description
1	onchainos security token-scan	Token risk / honeypot detection (all chains)
2	onchainos security dapp-scan	DApp / URL phishing detection (chain-agnostic)
3	onchainos security tx-scan	Transaction pre-execution security (EVM + Solana)
4	onchainos security sig-scan	Message signature security (EVM only)
5	onchainos security approvals	Token approval / Permit2 authorization query (EVM only)

Reference Loading Rules (MANDATORY)

Before executing ANY security command, you MUST read the corresponding reference document from skills/okx-security/references/. Do NOT rely on prior knowledge — always load the reference first.

User intent	Read this file FIRST
Token safety, honeypot, is this token safe, 代币安全, 蜜罐检测, 貔貅盘	references/risk-token-detection.md
DApp/URL phishing, is this site safe, 钓鱼网站	references/risk-domain-detection.md
Transaction safety, tx pre-execution, signature safety, approve safety, 交易安全, 签名安全	references/risk-transaction-detection.md
Approvals, allowance, Permit2, revoke, 授权管理, 授权查询, 风险授权	references/risk-approval-monitoring.md

When a workflow involves multiple commands (e.g., token-scan then tx-scan), load each reference before executing that command.

Integration with Other Skills

Security scanning is often a prerequisite for other wallet operations:

• Before wallet send with a contract token: run token-scan to verify token safety
• Before wallet contract-call with approve calldata: run tx-scan to check spender
• Before interacting with any DApp URL: run dapp-scan
• Before signing any EIP-712 message: run sig-scan

Use okx-agentic-wallet skill for the subsequent send/contract-call operations.