What Exists In The Public Experiment
| Area | Current state |
|---|---|
| Top-100 catalog | Public chart snapshots, chart pages, matrix rows, and per-chart caveats. |
| Variants | Base variants and derived ConfigHub variants for selected chart paths. |
| Delivery evidence | Argo OCI evidence for committed paths, plus live parity and observation lanes. |
| Sceptic tests | Known gaps, fuzzing, dry-run checks, drift findings, and refusal boundaries. |
What This Points Toward
| Idea | What would make it useful |
|---|---|
| Private catalogs | Bring private charts, wrapper charts, and customer overlays into the same model. |
| Fleet operations | Promote, patch, scan, and observe many variants with a clear blast radius. |
| Reverse reconcile | Accept an authorized live fix back into desired state only when policy and round-trip checks allow it. |
| AI agents | Use AI to propose app and ops changes while ConfigHub keeps evidence, gates, and rollback records. |
| AICR, NIM, and stacks | Apply the same recipe, variant, and evidence model to larger AI and platform stacks. |
Guardrails
Do not describe planned ideas as shipped behavior. The public experiment uses evidence-backed words: pass, watch, blocked, refused, not applicable, and planned.
That honesty is part of the product story: expert users need to know which claims are strong today and which still need more work.