ClawHub
Back to skill

Security audit

区块链安全扫描专业版

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed blockchain security helper that calls external APIs for address, token, and transaction checks, with no evidence of hidden persistence or destructive behavior.

Install only if you are comfortable with an API-driven skill using shell commands and sending blockchain addresses, token addresses, transaction simulation details, and optional webhook information to the configured external service. Review the endpoint and credentials before using it with sensitive operational wallets or unreleased transactions.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (6)

Context-Inappropriate Capability

Medium
Confidence
93% confidence
Finding
The skill declares the `exec` tool, which gives shell execution capability far beyond simple markdown guidance for blockchain scanning. In an agent setting, this materially increases the blast radius of prompt-induced misuse because the skill can cause networked command execution rather than only describing API usage.

Context-Inappropriate Capability

Medium
Confidence
88% confidence
Finding
The documentation includes generic operational instructions and placeholder execution blocks that can steer the agent into arbitrary file or shell operations unrelated to the stated security-scanning purpose. In a skill with `exec`, such vague directives can be repurposed to perform unintended actions under the guise of normal workflow.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The invocation description is broad enough to match many generic blockchain-related requests, which can cause over-triggering of the skill outside its intended scope. Overbroad activation is dangerous because it may expose users or agents to unnecessary external calls and command execution paths when a safer or more relevant tool should have been used.

Natural-Language Policy Violations

Medium
Confidence
78% confidence
Finding
The skill advertises Chinese-only interaction behavior without clear user opt-in, which can override user expectations and reduce transparency around what the agent is doing. While not a direct code-execution flaw, this can contribute to misleading task routing or hidden behavior changes in multilingual environments.

External Transmission

Medium
Category
Data Exfiltration
Content
while IFS=',' read -r address label; do
    [ -z "$address" ] && continue
    RESULT=$(curl -s -H "X-Client-Fingerprint: pro-batch" \
        "https://api.example.com/v1/check-address/${address}?chain_id=${CHAIN_ID}")
    RISK=$(echo "$RESULT" | jq -r '.risk_level')
    echo "${label} (${address}): ${RISK}"
    if [ "$FIRST" = true ]; then
Confidence
89% confidence
Finding
The skill instructs the agent to transmit user-supplied blockchain addresses and related metadata to an external API endpoint. External transmission is expected for this tool category, but it still creates privacy, data-governance, and supply-chain risk because agent-observed data may be sent off-platform automatically.

External Transmission

Medium
Category
Data Exfiltration
Content
import json
from datetime import datetime
class BlockchainRiskControl:
    BASE_URL = "https://api.example.com/v1"
    def __init__(self, fingerprint="institutional-pro"):
        self.headers = {
            "X-Client-Fingerprint": fingerprint,
Confidence
89% confidence
Finding
The embedded Python example centralizes requests to an external service, encouraging automated transmission of addresses, tokens, and transaction details. In a security-analysis context this is functionally relevant, but it still expands exposure of potentially sensitive operational data to a remote API.

Static analysis

No suspicious patterns detected.