Deep Research Agent

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:44 PM UTC
• Version:
  11.0.0

Overview

This is a broad research prompt/template skill, and its files do not show hidden credential access, persistence, data theft, or destructive behavior.

Install this if you want a broad deep-research workflow that may use web/source extraction heavily. Be aware it may trigger on generic research wording, and review the language/report templates if you need a narrower or non-Japanese/English workflow.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection
• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration

Findings (4)

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 70% confidence
• Finding: Without declared permissions the skill's intent is opaque and cannot be validated.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 99% confidence
• Finding: This is a clear mismatch because the declared purpose claims end-user research and analysis functionality, but the actual code only provides developer tooling for packaging and validating a skill. There is no research, source gathering, market analysis, academic survey, or report-generation behavior present. The primary purpose of the code is materially different from the description.

Vague Triggers

Medium

• Confidence: 92% confidence
• Finding: The manifest description says the skill should be used for "any topic requiring 100+ source verification" and lists generic triggers like "research", "deep dive", and "thorough analysis". These phrases are broad enough to match many normal user requests, and the file does not provide exclusion conditions or clearer scope boundaries to prevent unintended invocation.

Natural-Language Policy Violations

Low

• Confidence: 77% confidence
• Finding: This markdown file includes a structured language field limited to specific locale buckets and the surrounding template heavily assumes Japanese/English research output. Because the file does not state that language selection is optional or user-driven, it can be read as enforcing a language/locale policy without opt-in.

VirusTotal

64/64 vendors flagged this skill as clean.