Security audit

AWS Global Search

Security checks for vulnerabilities and agentic risk

Overview

This AWS lookup skill is mostly a documentation helper, but it includes copyable examples that can create billable AWS resources without clear safety warnings.

Review before installing. Use it for AWS documentation and pricing lookup, but do not paste real AWS access keys into examples or run the EC2/Lambda create commands unless you intentionally want to modify an AWS account; use least-privilege credentials, understand costs, and plan cleanup first.

A.I.G

Vulnerability Patterns

Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access

Findings (6)

Missing User Warnings

Medium

Confidence: 94% confidence
Finding: The EC2 CLI and Boto3 snippets are runnable resource-creation examples presented without warnings that they will launch billable infrastructure and alter a live AWS account. In agent contexts, users may execute copied commands verbatim, causing unintended spend and exposure from misconfigured security groups, keys, or networking.

Context-Inappropriate Capability

Medium

Confidence: 95% confidence
Finding: This no-login information-retrieval skill includes credentialed EC2 creation examples that require AWS access keys and perform state-changing actions. Even with placeholders, the content normalizes supplying secrets and executing billable operations in a context where users would reasonably expect read-only assistance.

Description-Behavior Mismatch

Medium

Confidence: 93% confidence
Finding: The examples materially expand a read-only AWS product/documentation lookup skill into operational guidance for creating resources and designing deployments. This scope drift can cause downstream agents or users to treat the skill as authorized for cloud changes, increasing the chance of unintended account modifications, spend, and unsafe automation.

Context-Inappropriate Capability

Medium

Confidence: 96% confidence
Finding: The Lambda example teaches authenticated API usage to create live cloud resources, which exceeds the declared capability of querying public AWS information. Such instructions can be copied into real environments, leading to unintended deployments, privilege misuse, and unexpected charges.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The Lambda examples provide complete create-function commands and SDK code without warning that they modify the user's account and can incur ongoing charges or security risk if roles and environment settings are misused. This is especially risky because the skill is framed as public information lookup, not account administration.

Context-Inappropriate Capability

Low

Confidence: 84% confidence
Finding: The manifest frames the skill as a source-querying tool for AWS product information, documentation, features, and pricing. This example demonstrates end-to-end solution architecture recommendation across multiple AWS services with design choices, cost modeling, and security best practices, which is advisory solution design rather than straightforward product lookup.

Static analysis

No suspicious patterns detected.