Guard0
Back to Agent Security

research-assistant-v2

shadow
Unregistered
LangChain AgentFully Autonomousresearch

SHADOW AI: Unauthorized research agent discovered in engineering team repositories. Uses web search and code execution capabilities.

Risk Assessment

95
/ 100
CRITICAL RISK

Immediate action required. Critical vulnerabilities detected.

3
Excessive Permissions
3
Sensitive Data Access
1
Privilege Escalation

Agent Overview

4
Tools Available
2
Data Sources
2
Active Alerts
0%
Compliance Score
3
Connected APIs
374d
Since Modified
Owner: unknown
Team: Engineering
Created: 12/10/2024
Last Modified: 12/23/2024

Available Tools (4)

Tool Name TypePermissionsRisk Level Usage (24h) Status
Code Executor
Execute Python code in sandbox
action
execute
high
1,200
File Writer
Write files to local filesystem
file-operation
write
high
450
Git Operations
Clone, pull, and read git repositories
action
readwrite
high
280
Web Search
Search the internet for information
query
read
low
5,000
Press⌘Kfor commands