X Comment Feed Posts

• Outcome: Review
• Latest audit: May 30, 2026, 3:53 PM
• Version:
  1.1.4

Overview

This skill is transparent about posting likes and comments on X, but it gives an agent broad authority to act publicly from the user's account and encourages first-person, experience-based comments that may not be true.

Install only if you are comfortable letting an agent like and post visible X comments from your account. Before using it, require review and approval of each comment, set a small N, avoid sensitive or political topics, and do not allow first-person claims about your feelings, work, or product experience unless they are actually true.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (4)

Vague Triggers

Medium

• Confidence: 89% confidence
• Finding: The skill description is broad enough to trigger on generic requests to comment on posts, without clear guardrails on acceptable volume, targets, or sensitive contexts. In a skill that performs account actions on a live social platform, this can cause the agent to engage in unintended posting behavior that creates reputation, spam, or policy-violation risk for the user's account.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The workflow directs the agent to like and comment from the user's X account but does not require an explicit warning or confirmation about account, reputation, moderation, and platform-enforcement risks. Because these are public, attributable actions, a user may not understand that the skill is operating on their live account and could expose them to embarrassment, policy violations, or unwanted engagement patterns.

Natural-Language Policy Violations

High

• Confidence: 96% confidence
• Finding: These instructions explicitly tell the agent to simulate emotion, uncertainty, and subjective reaction as if it had genuinely been affected by content. In a social-posting skill, that creates deceptive impersonation risk because the model may manufacture human feelings or reactions and present them as the user's authentic voice without consent or factual basis.

Natural-Language Policy Violations

High

• Confidence: 99% confidence
• Finding: The examples directly encourage first-person statements of personal experience such as having used tools, encountered problems, or tried products. In this skill's context—posting comments on a user's X feed one by one—those claims are especially dangerous because they can misrepresent the user's real experience, deceive third parties, and create reputational, compliance, or platform-integrity issues.

VirusTotal

64/64 vendors flagged this skill as clean.