Hi everyone! I'm Amrit, a CS student at IIIT, currently based in Tokyo. I work mostly in C++, Python, and TypeScript, with experience in backend systems and infrastructure.
I've been going through the 2026 ideas list and I'd like to propose an original idea: NixOS Configuration Drift Detector & Reconciler.
NixOS guarantees reproducible builds and atomic activation, but it does not continuously verify that runtime state remains aligned with the declared configuration. In practice, systems can diverge from their intended state due to manual interventions, mutable state changes, or runtime inconsistencies.
The core idea is to build a NixOS module + CLI tool that detects and reports configuration drift between the declared NixOS configuration and the live system state.
Initial scope could include:

Comparing systemctl list-units output vs declared systemd.services
Detecting firewall rules that differ from the configured networking.firewall state
Checking user accounts against users.users
Verifying /etc contents against activation outputs
Identifying services running with mismatched derivations or binaries

The tool would generate structured drift reports with severity levels and clear explanations. A reconciliation mode could optionally restore safe classes of drift or generate suggested configuration patches.
Validation would be done using NixOS VM tests to simulate controlled drift scenarios and ensure correctness.
Effort: Medium (175h) to Large (350h), depending on how far reconciliation and diagnostics are taken.
I'd really appreciate feedback on whether this aligns with current ecosystem priorities and whether anyone would be interested in mentoring or refining the scope. Happy to draft a more detailed design document or open a PR to the ideas list.
GitHub: https://github.com/arkhamknight25