ClawHub
Back to skill

Security audit

Aws Agentcore Langgr

Security checks for vulnerabilities and agentic risk

Overview

This is an AWS deployment guide, but its read-only declaration, broad activation terms, and under-scoped cloud command guidance make it something users should review before installing.

Install only if you want an agent to advise on AWS AgentCore/LangGraph deployment. Before allowing any command execution, verify the AWS account, region, resources, credentials, and whether a command will create or delete cloud infrastructure; do not let it run broad or automatic cloud actions from generic prompts.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Intent-Code Divergence

High
Confidence
95% confidence
Finding
The manifest exposes only a read tool, but the documentation instructs users to perform deployment, command execution, API calls, and write-capable cloud actions. This capability mismatch can mislead users and host agents about the skill's effective behavior, increasing the risk of unauthorized or unexpected high-impact operations if the surrounding agent environment grants broader permissions than declared.

Description-Behavior Mismatch

Medium
Confidence
85% confidence
Finding
The skill is presented as a lightweight deployment helper, but the body expands into generic automation including file processing, external API integration, and command execution. That broader scope can cause over-trust and unsafe invocation in contexts where users or orchestrators expect a narrowly scoped assistant, making accidental misuse more likely.

Vague Triggers

Medium
Confidence
83% confidence
Finding
The trigger keywords are very broad terms like 'aws', 'agentcore', and 'skill', which can cause the skill to activate in many unrelated conversations. Over-broad activation is dangerous here because the skill discusses command execution, deployment, and cloud operations, so accidental invocation could steer an agent toward risky actions in the wrong context.

Static analysis

No suspicious patterns detected.