ClawHub
Back to skill

Security audit

alibabacloud-mining-attack-diagnosis

Security checks for vulnerabilities and agentic risk

Overview

This is mostly a read-only Alibaba Cloud mining-investigation skill, but it needs Review because it can automatically use local cloud credentials for broad account security reads and probes unrelated alert details when no mining alert exists.

Install only if you want this skill to run Alibaba Cloud Security Center investigations from your local aliyun CLI profile. Use a dedicated read-only RAM role matching the documented permissions, avoid broad default profiles, and be aware that a clean mining result may still read one unrelated alert detail for workflow completeness.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill explicitly instructs the agent to execute a Python entry script, use the aliyun CLI profile, make networked API calls, write reports to disk, and log metadata. Those are real code-execution capabilities, yet the static finding says no permissions are declared, which creates a dangerous mismatch between apparent and actual power. In a security-analysis skill, this is especially risky because shell, network, environment, and file-write access could expose credentials, exfiltrate cloud data, or operate on the wrong account if the runtime trusts undeclared capabilities.

Description-Behavior Mismatch

Medium
Confidence
85% confidence
Finding
When no mining alerts exist, the script intentionally probes detail APIs using a general non-mining alert to keep the workflow complete. This causes the skill to access and process unrelated alert data outside the user’s stated mining-diagnosis purpose, creating unnecessary exposure of potentially sensitive security-event details and violating data-minimization expectations.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The trigger list contains broad terms such as 'miner', 'mining', 'risk assessment', and 'Security Center', which can match benign or unrelated user requests. Over-broad routing can cause this skill to activate unexpectedly and push the agent toward shell execution, cloud API access, and report generation in contexts where the user did not intend a cloud investigation. Because this skill has operational capabilities, accidental invocation increases the chance of unnecessary access to sensitive environment and account data.

Static analysis

Detected: malicious.crypto_mining

Possible crypto mining behavior detected.

Critical
Code
malicious.crypto_mining
Location
scripts/_constants.py:27