Linux automation engineer building AI agents that are reliable and easy
to keep under control, with security and confidentiality by design.
- Automate technical processes (Linux, Ansible, shell)
- Install, configure, and program AI agents, and develop plugins
- Contribute to open source, focused on security and secret handling
Building agents hardened by default: no access to secrets, read-only self-config, sensitive-value redaction, deterministic audit.
- hermes-agent-env-secret-redactor - a plugin masking exact secret values from tool output
- hermes-agent-news-fr - a daily digest of Hermes Agent news, in French
- tit4nium.fr - writing on agents, LLMs, and security

