great_cto

v2.41.0 Feature

This release adds 3 notable features for engineering teams evaluating rollout.

Published 1mo AI Agents & Assistants

View tool

✓ No known CVEs patched

Read the diff → Tool health → What is this tool? →

✓ No known CVEs patched in this version

Topics

agentic-coding claude-code-plugin claude-code-skills claude-code-subagents code-review cto

+2 more

multi-agent sdlc

Summary

AI summary

Autopilot flow execution now supports live FHIR and code‑set connectors with a stub/live dispatcher.

Changes in this release

Type	Severity	Summary	CVE
Feature
Feature	Medium	Connector runtime adds stub and live modes with lazy adapter registry. Connector runtime adds stub and live modes with lazy adapter registry. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Medium	Live FHIR/EHR connector reads real clinical notes from HAPI sandbox or configured endpoints. Live FHIR/EHR connector reads real clinical notes from HAPI sandbox or configured endpoints. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Medium	Live code‑sets connector performs ICD‑10‑CM / HCPCS lookup via NLM Clinical Table Search Service; flags CPT as needing licensed service. Live code‑sets connector performs ICD‑10‑CM / HCPCS lookup via NLM Clinical Table Search Service; flags CPT as needing licensed service. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Medium	Flow-runner executes flow steps, runs connector agents, and pauses at human checkpoint gate:coding-signoff. Flow-runner executes flow steps, runs connector agents, and pauses at human checkpoint gate:coding-signoff. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Low	CLI for flow-runner supports `--live` and `--full` flags to control execution mode. CLI for flow-runner supports `--live` and `--full` flags to control execution mode. Source: llm_adapter@2026-06-07 Confidence: high	—
Feature	Low	Adds 39 new unit tests, bringing total to 205. Adds 39 new unit tests, bringing total to 205. Source: llm_adapter@2026-06-07 Confidence: high	—
Performance	Low	End‑to‑end test demonstrates running two live connectors then halting at gate:coding-signoff. End‑to‑end test demonstrates running two live connectors then halting at gate:coding-signoff. Source: llm_adapter@2026-06-07 Confidence: high	—

Full changelog

Autopilots execute the flow — connector runtime + two live connectors (Phase 4)

The service-autopilot verticals (shipped in v2.40.0) stop describing a flow and start
executing it, with the first live connectors reading real data — no API keys.

Connector runtime (scripts/lib/connectors.mjs) — a call() dispatcher with stub | live
modes, a lazy live-adapter registry, hasLiveAdapter, and a CLI. Stub stays deterministic +
network-free; live hits a real adapter when one is registered, else falls back to stub.
Live FHIR/EHR connector (connectors/fhir.mjs) — a real FHIR R4 client; defaults to the
public HAPI sandbox (no auth), GREAT_CTO_FHIR_BASE/TOKEN for Epic/Cerner/athenahealth.
Reads real clinical notes.
Live code-sets connector (connectors/codesets.mjs) — real ICD-10-CM / HCPCS lookup +
validation via the NLM Clinical Table Search Service (free, no auth). CPT is AMA-licensed, so it
returns a clear "needs a licensed service" note rather than a wrong answer.
Flow-runner (scripts/lib/flow-runner.mjs) — the doing layer: walks a flow's steps, runs each
agent step's connectors, and pauses at a human checkpoint — the assistant↔autopilot boundary
enforced at runtime. CLI: flow-runner.mjs <vertical> [--live] [--full].

Proven end to end: flow-runner.mjs rcm --live runs two live connectors (real note via FHIR,
real ICD-10 via NLM) then halts at gate:coding-signoff. 39 new unit tests (205 lib total).

View diff on GitHub

Weekly OSS security release digest.

The CVE patches and breaking changes that affected production tools this week. One email, every Sunday.

No spam, unsubscribe anytime.

Share this release

Share on X Share on Bluesky

Track great_cto

Get notified when new releases ship.

About great_cto

Engineering-management layer of 34 specialist AI agents covering the full SDLC (architect, PM, senior-dev, reviewer, QA, security, devops, L3-support + 18 archetype-specific reviewers) with auto-detected archetypes and compliance gates (PCI-DSS, HIPAA, FedRAMP, GDPR, EU AI Act). Runs in Claude Code, Cursor, Codex CLI, Aider, and Continue via AGENTS.md + MCP. MIT.

All releases →

Related context

Related tools

Earlier breaking changes

v2.72.2 Removes Autopilot sidebar item and Operate topbar button from dev board.
v2.55.0 Hard-gates the Build board; redirects invite sessions to Operate for operators.
v2.43.0 Runtime now blocks autonomous execution of irreversible actions without prior human checkpoint.
v2.43.0 Adds reversible and blastRadius fields to every flow step.
v2.32.0 Removed AgentShield scanner and its CLI commands.