Security audit

TierList Maker

Security checks for vulnerabilities and agentic risk

Overview

This skill creates TierVibe tier lists and its file creation and browser handoff are mostly disclosed and tied to that purpose, though users should know board data may be stored locally in a generated HTML file.

Before installing, be aware that the skill creates local handoff files such as launcher.html and sometimes manifest.md or a .tiervibe.json file. Those files can contain your tier-list text and, for local image mode, embedded image data; delete them when no longer needed. Prefer opening the generated launcher yourself rather than allowing an agent to auto-open external URLs.

SkillSpector

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (3)

Missing User Warnings

Low

Confidence: 84% confidence
Finding: The skill directs the agent to create local files such as a manifest without first clearly informing the user that files will be written to disk. Unexpected filesystem writes can violate user expectations, leak sensitive content into persistent storage, or create artifacts in shared/work environments even if the files are not directly dangerous.

Missing User Warnings

Low

Confidence: 91% confidence
Finding: The workflow requires writing a local launcher.html file to disk as the normal operating path, but it does not present that disk write as an up-front consented action. Even though the file is meant to help the user, silently creating HTML files can surprise users, persist sensitive board content locally, and in some environments raise trust or policy concerns.

Missing User Warnings

Low

Confidence: 88% confidence
Finding: The example instructs the agent to automatically open an external website in the user's default browser without an explicit confirmation or clear warning at the moment of launch. While this is not remote code execution, it creates a consent and trust-boundary issue: the skill causes an external action and navigates the user to a third-party site with embedded data, which can surprise users and normalize unsafe auto-launch behavior.

Static analysis

No suspicious patterns detected.