Practical summary
Give developers a local diagnostics and interaction surface during evaluation.
OVERVIEW · Implemented
TUI & operator tools
Inspect local runtime state and explore session-based agent interactions.
MLAI Research · SEPTEMBER 2026 · 1 min read · overview
Status
Implementation status applies to the bounded source capabilities described here, not to a hosted product or benchmark result.
Reviewed 2026-09-30
What it can help with
Give developers a local diagnostics and interaction surface during evaluation.
These are potential pilot applications, not claims of customer outcomes.
What the source implements
The agent REPL exposes session commands for model selection, profile and status inspection, context, history and reset. The diagnostics dashboard supports panes and deterministic one-shot output. A session-local SEA preference does not by itself prove durable evidence retrieval occurred.
Available today
The diagnostics dashboard and agent REPL are implemented.
How to evaluate it
Use one-shot diagnostics in automation and a terminal for interactive acceptance. Confirm the displayed data source and fallback disclosures before interpreting a pane as an observation of a live service.
Supporting research and guides
Read the accompanying archive entries: Using the local diagnostics dashboard and agent REPL. Each distinguishes implemented behavior from proposed applications and includes pinned implementation evidence.
Limitations
- Interactive editing and refresh require a supported terminal.
- A dashboard view is not a production health certification or proof of provider execution.
Supporting sources
- Agent REPL commands and session state
source · revision
80dfe079ebe7413a6815c2d564be1f0aaa901267 - Diagnostics dashboard and one-shot output
source · revision
80dfe079ebe7413a6815c2d564be1f0aaa901267