ClawHub
Back to skill

Security audit

AWS图代理

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent AWS AgentCore/LangGraph deployment guide, with powerful but disclosed file, shell, AWS credential, and deployment requirements.

Install only if you intend to use AWS Bedrock AgentCore/LangGraph workflows. Use least-privilege AWS roles or profiles, do not paste or print AWS secrets, review generated code before deployment, require explicit approval before launch/destroy or refund-like tool actions, and only use callback URLs you control and trust.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Vague Triggers

Medium
Confidence
82% confidence
Finding
The skill’s activation description is extremely broad, covering generic code generation, debugging, deployment, and development tasks. In a skill with `Read`, `Write`, `Edit`, and `Bash` capabilities, vague invocation criteria increase the chance the agent is selected in unrelated contexts and given unnecessary access to files, shell execution, AWS deployment flows, and sensitive project state.

Vague Triggers

Medium
Confidence
84% confidence
Finding
The invocation guidance tells users to call the skill with ‘necessary input parameters’ but does not define hard trigger constraints, approval boundaries, or safe operating modes. Because this is an executable skill with shell and file-modification tools, the lack of constraints can lead to overbroad autonomous actions, including code changes, local command execution, and cloud deployment steps without sufficient narrowing.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The README exposes a `callback_url` parameter for asynchronous completion notifications but does not warn that using it may transmit data to an external endpoint. In an agent/deployment skill, this creates SSRF and data-exfiltration risk if untrusted users can supply arbitrary callback URLs or if outputs may contain sensitive operational details.

Static analysis

No suspicious patterns detected.