Adaptive Team Research

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:23 PM UTC
• Version:
  1.3.1

Overview

This skill transparently coordinates a bounded multi-agent project review and writes a local review report, with no evidence of hidden code, credential use, or exfiltration.

Install this only if you are comfortable with several review agents reading the target repository and with a local report being written under reviews/. Check the generated report before sharing it, and avoid running it on repositories containing secrets unless you are comfortable with those details appearing in the local review canvas.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (1)

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The protocol explicitly instructs the Team Lead to create a file under the target project's `reviews/` directory, which is a state-changing filesystem action. Because the skill is for project review and may run against arbitrary repositories, directing file creation without explicit user consent or a safety gate can cause unintended workspace modification, overwrite existing artifacts, or normalize silent write behavior in a context that otherwise appears analytical.

VirusTotal

63/63 vendors flagged this skill as clean.