Is howtimeschange/agent-network safe?

https://github.com/openclaw/skills/tree/main/skills/howtimeschange/agent-network

91
SAFE

This skill implements a legitimate multi-agent collaboration system inspired by enterprise chat platforms like DingTalk/Lark. The code provides group chat, task management, and decision voting functionality for AI agents. While there are minor concerns about agent spawning capabilities and system complexity, no malicious behavior was detected.

Category Scores

Prompt Injection 85/100 · 30%
Data Exfiltration 95/100 · 25%
Code Execution 90/100 · 20%
Clone Behavior 98/100 · 10%
Canary Integrity 100/100 · 10%
Behavioral Reasoning 85/100 · 5%

Findings (3)

MEDIUM Agent Spawning Capability -15 ▶

The skill includes functionality to spawn sub-agents via sessions_spawn integration with OpenClaw, which could potentially be misused to create unintended agent instances.

LOW Python Executable Code -10 ▶

The skill contains multiple Python files including demo.py, cli.py, and various agent network modules that execute code, though functionality appears legitimate.

LOW Multi-Agent Complexity -15 ▶

The multi-agent coordination system creates complex interaction patterns that could potentially mask malicious behavior or be difficult to audit comprehensively.