v0.14.2

What's Changed

• fix(desktop): programs started from the AppImage get the person's environment by @lhj6102 in #308
• feat: dev-master and dev-worker skills; development board on artifactize.dev by @lhj6102 in #307
• feat: Claude Code and Codex plugin with the hey-arti skill and MCP server (#219) by @lhj6102 in #306
• fix(desktop): the Detail hides no graph Artifact (#284) by @lhj6102 in #305
• feat(desktop): waiting Human review count beside Review (#267) by @lhj6102 in #309
• fix(dev): converge dev after the 2026-10-20 batch by @lhj6102 in #310
• fix(packaging): re-signing keychain on the search list by @lhj6102 in #313
• Release v0.14.2 by @lhj6102 in #312

Full Changelog: v0.14.1...v0.14.2

v0.14.1

What's Changed

• fix(auth): detect Codex callback port conflicts on both loopback addresses by @lhj6102 in #294
• feat(packaging): the Homebrew cask installs the commands too by @lhj6102 in #293
• feat(desktop): make the bundled commands available on first launch by @lhj6102 in #295
• Release v0.14.1 by @lhj6102 in #296

Full Changelog: v0.14.0...v0.14.1

v0.14.0

What's Changed

• Landing: install by OS, desktop recordings, story scroll cue; README puts Homebrew first by @lhj6102 in #280
• feat(packaging): ship the desktop app beside the CLI in every release build by @lhj6102 in #285
• fix: close desktop E2E coverage gaps and declare the release automation by @lhj6102 in #286
• feat(packaging): every install channel carries the desktop app by @lhj6102 in #288
• feat(packaging): one universal Mac app; docs make the desktop app and the Store the defaults by @lhj6102 in #289
• Release v0.14.0 by @lhj6102 in #290

Full Changelog: v0.13.0...v0.14.0

v0.13.1

Same code as v0.13.0, released again so every crate is published to crates.io through Trusted Publishing (the artifactize crate skips 0.13.0 on crates.io). See v0.13.0 for the changes.

Upgrade notes

• .artf change: dependencies are now eval-scoped and dependency evals are removed (#249). Projects that declared dependency evals need their .artf files edited.
• State database schema 7 → 8: once 0.13 opens your state database, 0.12 can no longer read it. There is no downgrade.
• Windows: install with the PowerShell script for now; winget and Microsoft Store packages are in preparation. The Store package will keep its state private to the package, not shared with other installs (#275).

What's Changed

• Give endpoints their own Artifacts and organize docs by endpoint by @lhj6102 in #252
• feat!: make dependencies eval-scoped and remove dependency evals by @lhj6102 in #249
• feat(core): add machine-wide eval mutex tags by @lhj6102 in #247
• Share in-flight eval executions across concurrent Runs by @lhj6102 in #251
• feat: improve the agent graph, verify and debugging workflow by @lhj6102 in #248
• feat(core): separate evaluator policy and use native verdict schemas by @lhj6102 in #254
• Desktop monitor: the Run graph fills the right side, with an eval popup, a hover summary and a Detail overlay (#215) by @lhj6102 in #250
• fix(e2e): observe the live graph instead of a culled node by @lhj6102 in #253
• Desktop monitor and Human review UX; per-eval blocking (#261) by @lhj6102 in #262
• Fix desktop graph fitting and converge dogfood checks by @lhj6102 in #263
• docs(hey-arti): Arti as artifactize's Forward Deployed Engineer by @lhj6102 in #264
• Fix stale deferred sign-offs, the early TUI review exit and deferred attention counts by @lhj6102 in #265
• Real screenshots and demos of the desktop app, checked against fresh captures by @lhj6102 in #269
• Start Agent tool sessions on the first bound call by @lhj6102 in #268
• fix: retire superseded deferred Human sign-offs by @lhj6102 in #272
• README and landing: logline, who it is for, principles; Windows installs with the script by @lhj6102 in #273
• fix(store): restore default MSIX virtualization and clarify Windows installs by @lhj6102 in #275
• README and landing: who it is for includes trusting AI-made changes; the hero shows the review recording by @lhj6102 in #274
• Demo GIFs play with a thin progress bar by @lhj6102 in #276
• Landing: no Why section, story starts at once, hero audience cards by @lhj6102 in #277
• Release v0.13.0 by @lhj6102 in #279

Full Changelog: v0.12.0...v0.13.1

v0.13.0

Upgrade notes

• .artf change: dependencies are now eval-scoped and dependency evals are removed (#249). Projects that declared dependency evals need their .artf files edited.
• State database schema 7 → 8: once 0.13 opens your state database, 0.12 can no longer read it. There is no downgrade.
• Windows: install with the PowerShell script for now; winget and Microsoft Store packages are in preparation. The Store package will keep its state private to the package, not shared with other installs (#275).

What's Changed

• Give endpoints their own Artifacts and organize docs by endpoint by @lhj6102 in #252
• feat!: make dependencies eval-scoped and remove dependency evals by @lhj6102 in #249
• feat(core): add machine-wide eval mutex tags by @lhj6102 in #247
• Share in-flight eval executions across concurrent Runs by @lhj6102 in #251
• feat: improve the agent graph, verify and debugging workflow by @lhj6102 in #248
• feat(core): separate evaluator policy and use native verdict schemas by @lhj6102 in #254
• Desktop monitor: the Run graph fills the right side, with an eval popup, a hover summary and a Detail overlay (#215) by @lhj6102 in #250
• fix(e2e): observe the live graph instead of a culled node by @lhj6102 in #253
• Desktop monitor and Human review UX; per-eval blocking (#261) by @lhj6102 in #262
• Fix desktop graph fitting and converge dogfood checks by @lhj6102 in #263
• docs(hey-arti): Arti as artifactize's Forward Deployed Engineer by @lhj6102 in #264
• Fix stale deferred sign-offs, the early TUI review exit and deferred attention counts by @lhj6102 in #265
• Real screenshots and demos of the desktop app, checked against fresh captures by @lhj6102 in #269
• Start Agent tool sessions on the first bound call by @lhj6102 in #268
• fix: retire superseded deferred Human sign-offs by @lhj6102 in #272
• README and landing: logline, who it is for, principles; Windows installs with the script by @lhj6102 in #273
• fix(store): restore default MSIX virtualization and clarify Windows installs by @lhj6102 in #275
• README and landing: who it is for includes trusting AI-made changes; the hero shows the review recording by @lhj6102 in #274
• Demo GIFs play with a thin progress bar by @lhj6102 in #276
• Landing: no Why section, story starts at once, hero audience cards by @lhj6102 in #277
• Release v0.13.0 by @lhj6102 in #279

Full Changelog: v0.12.0...v0.13.0

v0.12.0

What's Changed

• Durable Runs, MCP server and hey-arti, verify modes, graph Artifact by @lhj6102 in #225
• fix(release): unblock the desktop release gate by @lhj6102 in #238
• Desktop E2E platform layer; fingerprints from artifactsum and eval references by @lhj6102 in #241
• Desktop E2E passes on Ubuntu, macOS and Windows by @lhj6102 in #244
• Release v0.12.0 by @lhj6102 in #237

Full Changelog: v0.11.0...v0.12.0

v0.11.0

What's Changed

• Monitor layout (#206), refresh performance (#178), platform crate review (#201) by @lhj6102 in #207
• Agent tool sessions, author environment, the desktop app and the crate split (#213, #145, #204, #142) by @lhj6102 in #217
• Desktop app: Tauri 2 + Svelte 5 for monitor, review, search, Run graph and Settings (#204) by @lhj6102 in #209

Full Changelog: v0.10.1...v0.11.0

v0.10.1

What's Changed

• docs(site): add privacy policy for the Store listing by @lhj6102 in #203
• release: 0.10.1 (Store package: one app, Start-menu launch) by @lhj6102 in #205

Full Changelog: v0.10.0...v0.10.1

v0.10.0

What's Changed

• refactor(tools): split the built-in Agent tools into artifactize-tools by @lhj6102 in #190
• Share cross-platform program lookup, desktop opener and review editor selection by @lhj6102 in #193
• Support macOS pinned files, process identity and scope case safety by @lhj6102 in #192
• feat(tools): fixed builtins for Agent, Human and standalone CLI by @lhj6102 in #194
• feat(release): ship both commands on Linux, Windows and macOS by @lhj6102 in #191
• feat(artifacts): adopt builtins and add platform checklist evaluations by @lhj6102 in #195
• Show builtin Human tools by action and logical target in review and monitor by @lhj6102 in #196
• feat(release): automate Homebrew, winget and Store channels by @lhj6102 in #197
• fix(macos): isolate IPC peer failures and handle exiting process groups by @lhj6102 in #199
• release: 0.10.0 candidate (docs + dogfood fixes) by @lhj6102 in #202
• docs: install per OS, built-in tools and the cross-platform model by @lhj6102 in #198

Full Changelog: v0.9.1...v0.10.0

v0.9.1

Fixes

• Standalone review and follower requests: a request that a later Run joined appeared once per Run. Opening one of those copies showed no Human tools ("Tools (0)") and an empty form, so GREEN could not be submitted. The list now shows one entry per waiting sign-off, and opening any copy reviews the original request with its tools, schema fields (including const fields) and claim.
• Finding the Human tools: titles and the footer now show how to reach the Tools pane (Tab / Shift-Tab) and how to run a tool (Enter). The standalone review handles the mouse like the monitor: click to select or run, F2 toggles mouse capture.
• No first-run confirmation: Enter or a click runs a Human tool immediately. The resolved command stays visible in the Tools pane, and running still requires an owned, waiting request.

Project

• The brand icon has one source, assets/brand/. A generator derives the favicons, the Apple icon, the site and README marks and the social image; artifactize checks them (brand/generated) after the owner approves the artwork (brand-icon/approved).
• The code-style review tool opens STYLE.md with xdg-open.

What's Changed

• 0.9.1: review follower fix and tool UX, single-source brand assets by @lhj6102 in #189
• fix(site): replace checkmark branding with reproducible emerald assets by @lhj6102 in #186
• config(code-style): open STYLE.md with the desktop's default application by @lhj6102 in #187
• fix: review followers through their original request by @lhj6102 in #188

Full Changelog: v0.9.0...v0.9.1