Social Post Generator Agent

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:29 PM UTC
• Version:
  1.0.0

Overview

This skill does what it advertises: it reads a user-provided article file or URL and prints social post drafts, with no evidence of hidden posting, credential theft, or persistence.

Install if you want a shell-based helper that reads content you provide and fetches URLs you provide. Do not pass private files, internal network URLs, localhost/cloud metadata endpoints, or URLs containing tokens unless you intend that runtime to read or request them.

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
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (5)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 89% confidence
• Finding: The skill advertises shell-based execution capability via scripts and a required binary (`curl`) but does not declare corresponding permissions. This creates a transparency and policy-enforcement gap: users or orchestrators may invoke the skill without realizing it can execute commands and make network requests, increasing the risk of unintended code execution pathways or unsafe deployment in restricted environments.

Missing User Warnings

Medium

• Confidence: 91% confidence
• Finding: The README states that the skill fetches content from URLs but does not clearly warn users that invoking it may cause outbound network access. In agent environments, undisclosed network egress can expose sensitive URLs, enable SSRF-style access to internal resources if inputs are not constrained, or violate execution policies where external communication must be explicit.

Missing User Warnings

Low

• Confidence: 90% confidence
• Finding: The script will automatically make an outbound network request when the input looks like a URL, but it does not clearly warn the user that remote content will be fetched. This can expose the user's IP address, trigger access to internal or sensitive endpoints if untrusted input is supplied, and create privacy or SSRF-style concerns in environments where the agent may have broader network reach than the user expects.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: The script makes outbound HTTP requests to any user-supplied URL with curl and gives no explicit warning or consent prompt before transmitting data to a remote host. In an agent/skill context, this can trigger unintended network access, expose internal-only URLs to probing (SSRF-style behavior), and leak usage metadata such as IP address and user agent to attacker-controlled servers.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: The script performs network requests to arbitrary user-supplied URLs without explicit notice, validation, or restrictions. In an agent or automation context, this can enable SSRF-style access to internal services, unexpected data egress, or contacting attacker-controlled infrastructure simply by passing a crafted URL.

VirusTotal

60/60 vendors flagged this skill as clean.