Maker From Photo

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:43 PM UTC
• Version:
  1.0.0

Overview

This is a cloud photo-to-video helper whose network use and token handling fit its purpose, with some input-scope disclosures users should notice.

Install only if you are comfortable sending selected photos, prompts, media URLs, and generated project data to mega-api-prod.nemovideo.ai. Avoid sensitive personal, internal, or proprietary media and avoid private/internal URLs unless you trust the provider's handling, retention, and deletion practices.

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 Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access

Findings (3)

Description-Behavior Mismatch

Medium

• Confidence: 96% confidence
• Finding: The skill is advertised as taking uploaded still photos, but the implementation explicitly supports arbitrary remote URLs and broad media types including video, audio, and other formats. This expands the trust boundary beyond the stated purpose, can enable unanticipated data ingress or fetching of third-party content, and increases the risk of SSRF-like abuse, policy bypass, or accidental processing of sensitive/non-photo media.

Vague Triggers

Medium

• Confidence: 90% confidence
• Finding: The routing rule sends essentially all unmatched requests into the SSE action path, creating an overly permissive command surface. That increases the chance the skill will process unrelated or unsafe prompts, trigger backend actions outside the intended photo-to-video workflow, or mishandle ambiguous user input in ways the user did not authorize.

Missing User Warnings

Medium

• Confidence: 95% confidence
• Finding: The skill sends user media and prompts to a cloud backend, but the user-facing description and getting-started guidance do not clearly foreground that data leaves the local environment. This is a privacy and consent problem: users may upload personal photos or sensitive prompts without realizing they are being transmitted to a third-party service for processing.

VirusTotal

63/63 vendors flagged this skill as clean.