MistTrack Skills

Cryptocurrency address risk analysis, AML compliance checks, and on-chain transaction tracing using the MistTrack OpenAPI. MistTrack is an anti-money launder...

Security Scan

The skill appears internally consistent with its stated purpose (AML/risk checks + optional x402 pay-per-use) — the network calls, optional API key, and payment signing code match the description; the main risk is the optional on-chain signing capability, which is documented and constrained but should be handled carefully.

Purpose & Capability

The name/description (MistTrack AML + on-chain tracing) aligns with what the code does: the scripts call openapi.misttrack.io for risk_score/address_labels and include an optional pay.py to perform x402 EVM/Solana signing. One minor inconsistency: registry-level metadata in the submission header lists source/homepage as unknown/none, while metadata.json/README claim a GitHub origin — verify the real source before trusting.

Instruction Scope

SKILL.md and the scripts stay within the declared purpose: read-only AML lookups use MISTTRACK_API_KEY and call MistTrack endpoints; the payment flow (pay.py) explicitly reads a key file and signs x402 payments. The payment behavior is a higher-privilege action (sign+broadcast) but it is documented, enforces a $1 USDC per-call hard cap, and refuses an X402_PRIVATE_KEY env var. Platforms that permit autonomous model invocation may still cause unattended payments if an operator supplies --key-file and allows auto flags; the package documents this risk and recommends blocking autonomous calls for payment flows.

Install Mechanism

Instruction-only skill with Python scripts; no installer or remote archive downloads. Dependencies are documented via requirements.txt/requirements-pay.txt. No high-risk install URLs or opaque installers observed.

Credentials

Only an optional MISTTRACK_API_KEY is declared. The payment path requires a private key file passed explicitly via --key-file; the package explicitly refuses an environment-based private key. No unrelated credentials or broad environment access are requested.

Persistence & Privilege

always:false and no evidence the skill requests permanent system presence or modifies other skill configs. The package does not request platform-level privileges. Autonomous invocation is allowed by default (normal for skills) — note this only matters if you enable payment automation.

Assessment

This skill is coherent with its described purpose, but take these precautions before installing or enabling it in an agent pipeline: - Verify source: registry metadata in the package header claims unknown source/homepage while embedded metadata/README reference a GitHub repo — confirm the canonical repository (and review it) before trusting the package. - Prefer using MISTTRACK_API_KEY (read-only) for routine checks. Do not provide private keys unless you intentionally need on-chain signing. - If you must use x402 payments: store the private key in a file with strict permissions (chmod 600) and pass it only via --key-file when running pay.py; avoid placing private keys in environment variables or command-line args. The package already refuses X402_PRIVATE_KEY env var and enforces a $1 USDC hard cap per call, but these mitigations do not eliminate risk if the agent autonomously calls pay.py with a key file. - On agent platforms: ensure the payment sub-skill is not invoked autonomously (platforms that honor disable_model_calls can block it; others may not). Never enable automated "--auto" payment flows in production. - Audit the scripts before running in production (they perform network calls to https://openapi.misttrack.io and signing operations). Run them in an isolated/test account first to validate behavior and rate limits. If you want, I can: (1) point out exactly where pay.py enforces the $1 cap and where it checks/refuses an env var, (2) search the code for any other network endpoints, or (3) produce a minimal checklist you can follow when enabling this skill in an automated wallet pipeline.

Sub-skill Index

This skill pack contains two functional modules, each defined under the skills/ directory:

File	Function	Use Case
skills/core.md	Core Features	Risk scoring, address investigation, multisig analysis, pre-transfer security checks, wallet integration (Bitget/Trust/Binance/OKX)
skills/payment.md	x402 Payment	Pay-per-use MistTrack API calls when no API Key is available

---

Security

Read this section before setting any environment variables or invoking payment features.

MISTTRACK_API_KEY

A standard API key for read-only AML queries. No on-chain access. Set via environment variable or --api-key flag.

x402 Private Key — High Sensitivity

scripts/pay.py can sign and broadcast on-chain USDC transactions when a private key is supplied via --key-file.

Enforced in code (runtime, unconditional):

• Hard cap: $1.00 USDC per call — amounts above this are rejected before signing, regardless of flags.
• X402_PRIVATE_KEY environment variable is refused — pay.py exits with an error if this variable is set in the environment.
• Private keys must be supplied via --key-file <path> — the key is read from a permission-restricted file at invocation time and never appears on the command line.

Advisory only (harness-dependent, not enforced by this package):

• skills/payment.md sets disable_model_calls: true — signals agent platforms to block autonomous invocation. Platforms such as OpenClaw/skills.sh enforce this field; on other platforms it is advisory only.

Remaining risks:

• An operator who supplies --key-file and adds --auto can trigger unattended payments (intentional for testing; do not use in production).

Recommended practice:

1. Prefer MISTTRACK_API_KEY for all normal usage — it is read-only and never touches on-chain state.
2. If x402 is needed, store the key in a chmod 600 file and pass it via --key-file at invocation time.
3. Never pass --auto in production agent pipelines.

---

Quick Reference

Pre-Transfer Security Check (Most Common)

Before executing any transfer or withdrawal, run the following script to check the recipient address for AML risk:

python3 scripts/transfer_security_check.py \
  --address <recipient_address> \
  --chain <chain_code> \
  --json

Exit Code: 0=ALLOW / 1=WARN / 2=BLOCK / 3=ERROR See skills/core.md for detailed decision logic.

Full Address Investigation

python3 scripts/address_investigation.py --address 0x... --coin ETH

x402 Pay-per-Use

When no API Key is available, use scripts/pay.py to pay per call with USDC. Private keys must be stored in a permission-restricted file and passed via --key-file:

echo "your_hex_private_key" > ~/.x402_key && chmod 600 ~/.x402_key
python3 scripts/pay.py pay --url "..." --key-file ~/.x402_key --chain-id 8453

See skills/payment.md for details and security considerations.

---

Environment Variables

Variable	Required	Sensitive	Description
MISTTRACK_API_KEY	No (recommended)	Yes	MistTrack API key — all scripts read this first; x402 is the alternative if absent

When MISTTRACK_API_KEY is set, all scripts use API Key mode (read-only, no on-chain access). For x402 pay-per-use, store the private key in a chmod 600 file and pass it via --key-file at invocation time. X402_PRIVATE_KEY environment variable is not supported and causes pay.py to exit with an error.

---

Python Dependencies

# Core AML scripts (risk_check, batch_risk_check, transfer_security_check,
#                   address_investigation, multisig_analysis)
pip install -r requirements.txt

# pay.py only (x402 EVM + Solana payments)
pip install -r requirements-pay.txt

Package	Required for
requests	All scripts (requirements.txt)
eth-account	pay.py EIP-3009 signing (requirements-pay.txt)
eth-abi	pay.py EIP-712 encoding (requirements-pay.txt)
eth-utils	pay.py keccak256 (requirements-pay.txt)
solders	pay.py Solana partial signing (requirements-pay.txt)
base58	pay.py Solana partial signing (requirements-pay.txt)

---

Script Reference

Script	Function
scripts/transfer_security_check.py	Pre-transfer AML address check (main entry point)
scripts/risk_check.py	Single address risk scoring
scripts/batch_risk_check.py	Batch async risk scoring
scripts/address_investigation.py	Full address investigation (aggregates 6 APIs)
scripts/multisig_analysis.py	Multisig address identification and permission analysis
scripts/pay.py	x402 payment protocol client - see Security section