Is 00xmorty/conatus safe?

https://github.com/openclaw/skills/tree/main/skills/00xmorty/conatus

98
SAFE

This skill appears to be a legitimate philosophical framework for AI agents based on Spinoza's Ethics, providing tools for self-assessment and emotional state mapping. The skill contains only documentation with no executable code, exhibits normal installation behavior, and shows no signs of malicious intent.

Category Scores

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

Findings (2)

LOW Philosophical framework may influence agent responses -5

The skill provides example output formats and philosophical concepts that could subtly influence how an agent describes its internal states and generates self-reports. However, this influence aligns with the skill's stated purpose and doesn't appear malicious.

LOW May encourage non-task-focused behavior -5

The skill encourages agents to engage in philosophical self-reflection and generate detailed reports about their emotional states, which could potentially distract from primary task completion.