ClawHub
Back to skill

Security audit

aws-infra

Security checks across malware telemetry and agentic risk

Overview

This AWS inspection skill is mostly aligned with read-only infrastructure review, but it also under-discloses local AWS credential/profile handling and broad exec/write capabilities.

Install only if you are comfortable letting the agent run AWS CLI commands and write local report files. Prefer AWS SSO or short-lived role credentials, avoid pasting secrets into commands or shared terminals, use per-command --region/--profile flags where possible, and keep exported AWS results out of repos and support bundles unless reviewed.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (5)

Intent-Code Divergence

Medium
Confidence
94% confidence
Finding
The skill is described as read-only, but it instructs users to run `aws configure set region us-west-2`, which mutates the local AWS CLI configuration. This is not a cloud-side destructive action, but it still changes operator state and can silently redirect subsequent commands to a different region, causing mistakes, confusion, or accidental use against the wrong environment.

Intent-Code Divergence

High
Confidence
98% confidence
Finding
The document claims outputs do not contain secrets, yet the MFA/profile examples instruct users to set `aws_access_key_id`, `aws_secret_access_key`, and `aws_session_token` directly via CLI configuration. This encourages handling and persistence of credential material in ways that may be exposed through shell history, logs, screenshots, shared terminals, or local config files.

Description-Behavior Mismatch

Medium
Confidence
86% confidence
Finding
The manifest and primary description position the skill as an AWS read-only inspection tool, but later sections broaden it into generic file processing, file writing, API integration, and sandboxed command execution. That capability expansion increases attack surface and may cause an agent or user to trust the skill more than warranted, enabling actions beyond the advertised scope.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The skill advises exporting AWS query results to local files without warning that those outputs can contain sensitive infrastructure metadata such as instance IDs, tags, bucket names, network topology, and audit-relevant information. In shared workstations, repos, CI artifacts, or support bundles, those files can become an easy source of reconnaissance or unintended disclosure.

Credential Access

High
Category
Privilege Escalation
Content
export AWS_PROFILE=production
aws ec2 describe-instances
# ...
# 配置多个Profile (在~/.aws/credentials中)
# [production]
# aws_access_key_id = AKIAXXXXXXXX
# aws_secret_access_key = YOUR_API_KEY
Confidence
97% confidence
Finding
~/.aws/credentials

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal

Static analysis

No suspicious patterns detected.