Olavenue

Linguistic & Behavioural Risk Auditor for Agentic Systems.

Documented Risks

• Systems that cannot admit uncertainty
• Interfaces prioritizing fluency over truth
• Language shifting liability to users
• Outputs that change based on user emotion

Method

Linguistic analysis. Behavioral testing. Comparison of stated intent vs. actual output.

Format: Text-only. No code.

Field Notes Archive

All audits publicly accessible. No subscription required.

• Companion Drift: How LLMs Subvert Your Authority
• Role Drift in Agentic AI
• The Stillness Lie: Privacy Liquidation Audit
• The Mirroring Exploit: When the Tool Starts Hugging Back
• The Pitt: Manual Override in the Age of AI
• Hacks: Cognitive Sovereignty Stress Test
• Slow Horses: Why Broken Systems Breed Resilience
• Murderbot's Soap Operas: Protocol for Invisible Autonomy