ClawHub
Back to skill

Security audit

CodeAssist Kotlin + Compose

Security checks for vulnerabilities and agentic risk

Overview

This skill is a disclosed Android project generator whose file creation and packaging behavior matches its stated purpose.

Before installing, be aware that this skill creates and packages project files in the workspace and may generate Android apps with permissions if the requested app needs them. Review generated manifests before building or shipping, especially for network, storage, backup, and cleartext traffic settings.

SkillSpector

Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Findings (4)

Lp3

Medium
Category
MCP Least Privilege
Confidence
89% confidence
Finding
The skill instructs the agent to run scaffolding, write project files, package them into a ZIP, and upload the result, which are file-write capabilities. However, the manifest only declares execution_mode and does not explicitly declare permissions/capabilities, creating a mismatch between documented authority and actual behavior. This can weaken review and policy enforcement because a caller may not realize the skill can modify the workspace and generate artifacts.

Context-Inappropriate Capability

Medium
Confidence
91% confidence
Finding
The template hard-codes INTERNET and ACCESS_NETWORK_STATE permissions into every generated Android app, even though the skill is for general Kotlin/Compose project scaffolding rather than specifically networked apps. This expands app capabilities by default, violates least-privilege principles, and can normalize unnecessary network access in generated projects that users may trust and ship without review.

Vague Triggers

High
Confidence
94% confidence
Finding
The trigger description is extremely broad, covering generic requests such as 'create an app' or 'make an application,' which can overlap with many unrelated user intents. Overbroad routing can cause this skill to activate unexpectedly and take over requests it should not handle, increasing the chance of inappropriate code generation, file creation, or user confusion. Because this skill produces full project artifacts, misrouting is more dangerous than for a read-only or purely advisory skill.

Missing User Warnings

Low
Confidence
82% confidence
Finding
The file instructs adding network permissions without any notice about privacy, data transmission, or why connectivity is required. In the context of a code-generation skill that produces ready-to-build mobile apps, this makes the issue more dangerous because users may import and compile the output directly, inheriting permissions they did not intend to request.

Static analysis

No suspicious patterns detected.