From a15425762e815d67b45bd1193fdedf3c489ed953 Mon Sep 17 00:00:00 2001 From: Bryan Helmkamp Date: Sat, 7 Mar 2026 11:02:03 -0500 Subject: [PATCH] Move Models API, Verification API, and streaming logs into accordion sections Co-Authored-By: Claude Opus 4.6 --- docs/changelog/2026-03-07.mdx | 22 ++++++---------------- 1 file changed, 6 insertions(+), 16 deletions(-) diff --git a/docs/changelog/2026-03-07.mdx b/docs/changelog/2026-03-07.mdx index b45469c18..6ce960e9a 100644 --- a/docs/changelog/2026-03-07.mdx +++ b/docs/changelog/2026-03-07.mdx @@ -3,14 +3,6 @@ title: "Models API, exe.dev sandboxes, and auto-install agent CLIs" date: "2026-03-07" --- -## Models API endpoint - -A new `GET /models` endpoint exposes the full LLM model catalog via the REST API. This is the same data available through `arc models list` on the CLI, now accessible to the web UI and API consumers. Results are paginated. - -```bash -curl -H "Authorization: Bearer $TOKEN" https://your-arc-server/models -``` - ## exe.dev sandbox provider (beta) Arc can now run agent stages inside ephemeral [exe.dev](https://exe.dev) VMs as an alternative to Daytona sandboxes. The new `arc-exe` crate manages VM lifecycle through SSH — a management plane connection to `exe.dev` for creating and destroying VMs, and a data plane connection to the VM itself for command execution and file I/O. @@ -24,18 +16,10 @@ environment = "exe" Agent CLI tools (like `claude` or `codex`) are now automatically detected and installed inside sandboxes at runtime when they're missing. This includes installing Node.js via NodeSource if needed. Previously, you had to build custom Dockerfiles that pre-installed every CLI tool your workflow used. -## Verification API restructured - -The verification API has been reorganized from a flat `/verifications` namespace into distinct resources: `/verification/criteria` and `/verification/controls`. The run sub-resource path is now singular: `/runs/{id}/verification`. - ## Updated model catalog The default OpenAI model has been upgraded from gpt-5.3 to **gpt-5.4**. New models added to the catalog include **gpt-5.3-codex-spark** and **claude-sonnet-4-6**. Model defaults are now centralized in the catalog rather than hardcoded across crates. -## Streaming stage logs for CLI agents - -Stage logs now update in real-time while CLI agents are working. Arc syncs `cli_stdout.log` and `cli_stderr.log` to the stage directory on each poll iteration, so you can see what an agent is doing before it finishes — useful for long-running stages. - **Verification API paths have changed.** `/verifications` is now split into `/verification/criteria` and `/verification/controls`. The run sub-resource moved from `/runs/{id}/verifications` to `/runs/{id}/verification`. @@ -46,7 +30,13 @@ To migrate: ## More + +- New `GET /models` endpoint exposes the full LLM model catalog via the REST API with pagination +- Verification API reorganized: `/verifications` split into `/verification/criteria` and `/verification/controls`; run sub-resource path is now `/runs/{id}/verification` + + +- Stage logs now stream in real-time while CLI agents are working, so you can see progress before a stage finishes - Sub-workflow context diffs no longer leak child engine-internal keys (`internal.*`, `graph.*`, `thread.*`) into the parent run