---
title: "Studio Q1 release — GrowthScaler"
released: "February 14, 2026"
channel: "Studio note"
status: "Live in production"
authors: "Studio team"
---

# Studio Q1 release

It exists to mark what shipped, when, and where it landed. The longer thesis pieces live under [Briefs](blog.html). This format is short on purpose.

Three studio engagements crossed production this quarter. Two are in the operate phase; one is in the architect-to-build transition. From those motions, the studio shipped three new skill files into the shared library and one new engagement-model variant. The shape of each is below.

2026-02-14

### Skill file · `account-research.md` v2 shipped.

Pulled into two active engagements. Replaces the v1 single-pass research run with a structured retrieval-plus-synthesis pattern. Halves the run cost; improves the trace by an order of magnitude.

2026-02-07

### Skill file · `demo-composer.md` released.

New surface. Composes a hypothesis-led demo from an inbound signal plus the company's deployed system topology. First two production runs landed against the original sales motion they were tuned for.

2026-01-30

### Skill file · `pricing-math.md` generalized.

The per-seat-to-consumption math we ran for the first engagement is now a generalized skill. Bring an existing per-seat schedule, get back a consumption equivalent with the revert path built in.

2026-01-22

### Engagement model · Operator-in-residence pilot.

New shape. A senior operator embedded 60–80% time, owning the motion until the client team takes it. Two pilots running. Will be formalized as a standard engagement in Q2 if both exit cleanly.

2026-01-09

### Evaluation harness · Anthropic Claude 4.5 model-layer absorbed.

The eval harness shipped with every engagement absorbed the latest model-layer upgrade without rewriting a single skill file. This is the contract: skill files survive model drift.

## What is queued for next quarter.

Three threads carry forward. **One:** the content-engine retrieval layer, currently in design across two engagements, ships as a generalized skill file pair in early Q2. **Two:** the pricing-math skill gets a second variant for outcome-pricing schedules, paired with the underwriting playbook the studio has been running internally. **Three:** the operator-in-residence pilot either graduates to a standard engagement or gets retired against the second pilot's exit criteria.

The next studio note ships in the second week of May.