Repair agent (evaluator-bounded)

Proposes fixes, accepted only if they hold on new cases.

Takes inDiagnosed failures, gold set
The agentRepair agent
Hands backFix proposal with held-out evidence

Checked against clinicians

Held-out clinician labels; checks for judge overfit.

Where a person decides

Human approval; canary and rollback; PCCP-style change record.

Works on

Scribes, Q&A, chat, voice.

On the roadmap. Tell us if you need it sooner. Reviewed .

The problem it solves

Fixing a prompt for one failure can break three others. Fixes need proof, not hope.

How it works

01

Propose a fix

The agent suggests a prompt, retrieval or routing change for a diagnosed cause.

02

Test on held-out cases

The fix is judged on cases it hasn't seen.

03

Accept only if it holds

It's accepted only if it improves the target and doesn't harm other scores.

An example, illustrative

Before

Allergy omitted in 23% of notes.

With the agent

Prompt v7: 14%, held-out. No regressions. Awaiting approval.

What it needs from you

A diagnosis and held-out cases from earlier layers.

Its limits

Independent evidence for automated repair is thin. A person approves every change.

Quick answers

What does the repair agent do?

Propose prompt, retrieval or routing fixes; accept only if held-out clinician-labelled results hold within judge error bounds.

How is the repair agent checked against clinicians?

Held-out clinician labels; checks for judge overfit.

Does it change production by itself?

Never. A named human approves each change.

What are held-out cases?

Cases kept aside so the fix can't be tuned to them.

What can the repair agent not do?

Human approval; canary and rollback; PCCP-style change record. No agent makes a clinical decision.

Get early access to the repair agent