v0.6.1

Open Science v0.6.1

A polish-and-reliability release: a redesigned dotted-mark app icon, session pinning in the sidebar, resilient downloads with range-resume and retry, a one-click error report for failed runs, and a round of Codex/ACP and packaging fixes on top of v0.6.0.

Open Science is an open-source, model-agnostic AI workbench for scientific discovery — a self-hosted desktop app pairing a plan-and-execute agent with a persistent, sandboxed compute kernel and durable project/session storage.

This is a patch release: no breaking changes. It refines the desktop shell and download reliability, completes the migration off the legacy Local Claude provider, and hardens Codex/ACP session and permission handling.

✨ Highlights

• Redesigned app icon. A new dotted-mark app icon across macOS, Windows, and Linux builds and the in-app branding. (#370, #356)
• Session pinning. Pin key sessions to keep them at the top of the sidebar. (#319)
• Resilient downloads. Model-runtime and asset downloads now resume from a byte range after an interruption, show transfer speed, and retry automatically on transient failures. (#343)
• Report an error on a failed run. A Report error button on failed runs captures the failure for quick reporting instead of leaving a dead end. (#345)
• Claude subscription via setup-token. The isolated Claude subscription provider (paste a claude setup-token) replaces the removed legacy Local Claude login, running under an app-owned CLAUDE_CONFIG_DIR. (#350, #354)

🚀 New Features

• Session pinning — pin and unpin sessions so important work stays at the top of the session list. (#319)
• Resilient download manager — range-resume, live speed display, and automatic retry for interrupted downloads. (#343)
• Report error on failed runs — surface and report a failed run's error in one action. (#345)
• Claude subscription (isolated) provider — add a Claude subscription with claude setup-token; the token is stored encrypted and used under an app-owned config dir, isolated from your ~/.claude/. (#350)

🔧 Improvements

• New dotted-mark app icon wired through the packaged builds, dock, and in-app assets. (#370, #356)
• The legacy Local Claude provider has been removed; previously stored Local Claude entries are dropped on upgrade in favor of the Claude subscription flow. (#354)
• Failed-run reporting only surfaces genuinely unknown failures, not expected/handled ones, cutting noise. (#362)
• Cmd/Ctrl zoom-in, zoom-out, and actual-size shortcuts are re-enabled in the desktop app. (#341)
• Release artifact filenames are prefixed with aipoch- for clearer, consistent naming across platforms. (#367)
• The .deb package uses the build's actual architecture instead of a hardcoded amd64. (#369, #368)

🐛 Bug Fixes

• Codex MCP permissions — use the correct Codex MCP permission option IDs, and collapse allow_always to a single session-scoped option so grants don't leak. (#377, #342)
• Skill resolution — the agent is nudged to skills by their slug id so it can resolve and invoke them reliably. (#373)
• Stale ACP connections — session creation is blocked on a stale connection during a deferred reconnect, avoiding failures against a dead transport. (#348)
• Reviewer MCP identity — hyphenated Claude MCP tool titles are accepted so the reviewer recognises its own verification tool. (#353)
• AI review pipeline — dispatched AI reviews are stabilized, their outcomes reliably published, and a specific reviewer can be selected for a dispatched run. (#379, #378, #361, #357, #358, #349)

📦 Install

Requirements: macOS 12+ (Apple Silicon or Intel), Linux x64, or Windows 10/11 x64. On first run the app's onboarding wizard checks your environment and can install and configure a Claude-compatible model backend for you. Once installed, the app updates itself in place.

Download the build for your platform from the Assets below:

Platform	File
macOS (Apple Silicon)	aipoch-open-science-0.6.1-mac-arm64.dmg
macOS (Intel)	aipoch-open-science-0.6.1-mac-x64.dmg
Linux	aipoch-open-science-0.6.1-linux-x86_64.AppImage / aipoch-open-science_0.6.1_amd64.deb
Windows	aipoch-open-science-0.6.1-win-x64-setup.exe

macOS — first launch. Release builds are Developer ID signed and notarized by Apple, so they open like any other app. (A .app you build yourself is not notarized — see README → Building From Source.)

Windows — first launch (unsigned build). No Authenticode certificate yet, so SmartScreen shows a bypassable "unrecognized app" prompt (More info → Run anyway).

Build from source instead:

npm install
npm run build:mac   # or: build:linux / build:win

🧭 What's in this release (maturity)

Aligned with the PRD and Roadmap. This is a patch release: it polishes the desktop shell (icon, session pinning, zoom shortcuts), download reliability, and Codex/ACP handling on top of v0.6.0's remote-compute and SDK work. No Roadmap phase changes status.

• ✅ Implemented: desktop shell (Electron + React + TS) with a tray/single-instance lifecycle, reachable as a desktop app, an optional localhost web UI (with project/session deep links and lifecycle sync), a headless backend mode, and an open-science CLI daemon plus a zero-dependency Node.js task SDK; ACP agent runtime on a pluggable framework backend (Claude Code default, with OpenCode and Codex as alternatives — Codex usable via an existing ChatGPT/Codex subscription login, and Claude via an isolated claude setup-token subscription) whose app-managed runtimes are installable, switchable, and uninstallable from Settings; an opt-in reviewer with a bounded fix loop; a persistent multi-kernel notebook runtime (warm Python/R/REPL kernels, app-managed conda environments with offline provisioning and bring-your-own interpreters, crash-recoverable run history, session .ipynb export); project + session persistence with a paginated, session-grouped Files tab and session pinning; artifacts MCP server; file renderers + notebook/session-notebook viewers + a full-screen preview mode + offline molecule viewer; multi-provider config (built-in vendors Anthropic (Claude), DeepSeek, Zhipu AI (GLM), GLM Coding Plan, Kimi (Moonshot), MiniMax, StepFun, Xiaomi MIMO, SenseNova, Volcengine Ark, plus the OpenRouter aggregation gateway, custom gateways, and Claude/Codex subscription logins) with per-model image-input metadata and per-model reasoning-effort control; desktop task notifications; remote SSH compute hosts as notebook execution targets; resilient range-resume downloads; signed-and-notarized macOS packaging plus Windows/Linux, with in-place auto-update and a configurable data-storage location; file-based agent skills; 24 built-in connectors plus custom MCP servers; a guided onboarding wizard; and per-conversation approval profiles.
• 🚧 Partial: R remains managed-only and environment snapshotting/export is not built; the CLI/SDK surface is a first entry point and task-automation layer, not yet a fully scriptable/embeddable orchestration SDK; provider choice is still gated on the selected framework's endpoint compatibility rather than one unified gateway; remote compute covers SSH hosts for notebook runs but not Slurm/cloud-GPU submission or sub-agent fan-out; skills are local and file-based (no shared/public commons yet); the reviewer verifies a turn's own record but does not yet apply domain-specific citation/units/statistics rubrics.
• 🗺️ Roadmap (not in this build): public skills commons and savable specialist roles (rest of Phase 3); a fully model-agnostic gateway with unified per-agent routing and a full CLI/SDK entry point (rest of Phase 1); artifact versioning + provenance chain, environment snapshotting/export, and specialist sub-agents (rest of Phase 2); Slurm/cloud-GPU remote compute, sub-agent fan-out, and the rest of the security stack — network sandbox, credential vault, scoped permission tiers (rest of Phase 4).

🐢 Known Limitations

• R is managed-only. A bring-your-own R interpreter path is not built yet.
• No environment snapshotting or export yet. Phase 2 work.
• Remote compute is SSH-only. Slurm, cloud GPU, and sub-agent fan-out are not built yet.
• Provider choice is per backend, not one unified gateway. Claude Code reaches Anthropic-compatible endpoints; OpenCode and Codex add OpenAI-compatible / Responses ones.
• The task SDK is a first-generation surface. Task creation, polling, and artifact fetch work; a broader programmatic orchestration API is still open Phase 1 work.
• Switching backends can't carry a live session over. Open conversations replay to the new backend, but in-flight tool state is not migrated.
• Skills are local only. No shared/public commons, cross-machine forking, or user-facing version pinning yet.
• The reviewer is opt-in and record-scoped.
• Windows builds are unsigned. SmartScreen prompt on first launch. (macOS is notarized.)
• No provenance chain / artifact versioning yet. Phase 2.
• No local GPU compute backend.
• No multi-user / real-time collaboration.

🙏 Acknowledgements

Thanks to @ewen-poch, @imjszhang, @wen2zhou, @roxi3906, @GiaoLee, and everyone in Discord, X, and Discussions.

---

What's Changed

• fix(main): re-enable Cmd/Ctrl zoom in/out and actual-size shortcuts by @ewen-poch in #341
• fix(acp): collapse Codex MCP allow_always to single session-scoped option by @ewen-poch in #342
• fix(acp): block session creation on stale connection during deferred reconnect by @ewen-poch in #348
• fix(reviewer): accept hyphenated Claude MCP titles by @ewen-poch in #353
• chore(branding): replace app icon with dotted mark by @ewen-poch in #356
• ci(review): gate fork PR review and externalize review models by @ewen-poch in #349
• fix(review): restore AI review execution by @ewen-poch in #357
• feat(net): add resilient downloads with range resume, speed display, and retry by @ewen-poch in #343
• feat(workspace): add Report error button for failed runs by @ewen-poch in #345
• fix(review): restore working reviewer actions by @ewen-poch in #358
• feat(sidebar): add session pinning by @roxi3906 in #319
• feat(agent): add claude-isolated provider via setup-token by @ewen-poch in #350
• refactor(agent): remove the 'Local Claude' provider by @ewen-poch in #354
• build(packaging): prefix artifact names with aipoch- by @ewen-poch in #367
• fix(packaging): hardcode amd64 arch in deb artifactName by @ewen-poch in #368
• feat(branding): update app icon to dotted mark design by @ewen-poch in #370
• fix(packaging): use ${arch} instead of hardcoded amd64 in deb template by @ewen-poch in #369
• ci(review): allow selecting a dispatched reviewer by @ewen-poch in #361
• fix(acp): nudge skills by slug id so the agent can resolve them by @wen2zhou in #373
• fix(acp): use Codex MCP permission option IDs by @ewen-poch in #377
• ci(review): ensure AI review outcomes are published by @ewen-poch in #378
• fix(workspace): report only unknown run failures, not expected ones by @ewen-poch in #362
• ci(review): stabilize dispatched AI reviews by @ewen-poch in #379
• chore(release): bump version to 0.6.1 by @ewen-poch in #382

Full Changelog: v0.6.0...v0.6.1