Meldwork V1.0.5 — General Agent workspace prerelease

Meldwork V1.0.5

Meldwork V1.0.5 makes the local General Agent workspace more reliable across readiness checks, group execution, recovery, and result adoption.

Improvements

• Keeps installed Agents visible when a credential or capability probe fails, while isolating the failed probe and refreshing readiness safely.
• Preserves native shell, proxy, CA, Provider, and model-region settings during detection and refresh.
• Routes peer requests through explicit Agent decisions instead of treating prose mentions or media words as hidden commands.
• Separates task completion from individual invocation success and human adoption, so one healthy result remains usable when another participant fails.
• Resumes clarification gates after restart, rejects conflicting replies, and prevents duplicate responses from replaying calls.
• Serializes workspace writers, recovers interrupted writes through bound approvals, retains partial results, and preserves acknowledged session context.
• Keeps completed and stopped work visible after restart, including healthy contributions from a partially failed group.
• Updates workspace preference, sidebar, titlebar, and Pi capability-probe behavior.

Verification

• Frontend tests: 343/343 passed.
• Desktop tests: 1,562/1,562 passed.
• Deterministic evaluation: 6 cases and 18 results passed.
• Web and desktop builds, packaging, ZIP integrity, and deep code-signature verification passed.
• Packaged Apple silicon macOS acceptance passed for group execution, cancellation, and restart recovery.

Distribution

• Meldwork-0.1.5-arm64.dmg
• Meldwork-0.1.5-arm64.zip
• SHA256SUMS-0.1.5.txt

The preview artifacts are ad-hoc signed and are not signed or notarized with an Apple Developer ID. macOS may require Open Anyway on first launch. Live behavior remains dependent on the installed Agent CLI version, authentication, Provider, and capabilities.

Meldwork V1.0.4

This prerelease refactors group conversations around natural Agent discussion and task-scoped native session continuity.

Improvements

• Sequential discussion mode runs one Agent at a time in configured CLI order and records every round.
• Auto Discussion starts with concurrent proposals, then routes later turns from exact final-line @agent mentions. One mention runs one Agent; multiple mentions run concurrently.
• Natural discussion prompts let Agents exchange views before work and review instead of forcing every turn into a review-only role.
• Native Agent sessions are reused across rounds and restarts within one group task.

Reliability

• Durable checkpoints recover completed turns without duplicate reruns.
• Stale operation and snapshot bindings are rejected before late results can be committed.

Verification

• Frontend tests: 320/320 passed.
• Desktop workspace tests: 515/515 passed.
• Frontend Web and desktop renderer builds passed.
• DMG checksum verification, ZIP integrity, packaged app code-signature verification, and git diff --check passed.

Distribution

• Apple silicon macOS only: DMG and ZIP are included with SHA256SUMS.txt.
• The app is ad-hoc signed, not Apple Developer ID signed or notarized. macOS may require Open Anyway on first launch.
• Live Agent behavior depends on the installed CLI version, authentication, Provider, and declared capabilities.

Meldwork V1.0.3

Meldwork V1.0.3

This prerelease makes concurrent group conversations more predictable and keeps useful Agent work durable when a run is stopped.

Improvements

• Reply with the whole group by default: In Concurrent Responses, sending without an explicit Agent selection now invokes every Agent in the group. Scheduler limits may queue members, but no selected group member is removed from the batch.
• Cleaner conversation composition: Agent, Skill, and knowledge mentions now flow inline with the message, user metadata and Agent logos are easier to scan, Markdown formatting is rendered consistently, and mention menus follow the active selection while scrolling.
• Reliable conversation positioning: New messages return to the active Agent loading state, while the animated return-to-bottom control remains available when reading older content.
• Quieter unlimited mode: No round limit now uses the normal composer surface instead of a special background, border, shadow, or container breathing animation. The infinity cue and real phase feedback remain visible.

Reliability Fixes

• Completed and streamed concurrent replies remain visible and durable when the user stops a batch.
• Stopped attempts are recorded as interrupted, late results are rejected, and batch commits resume idempotently after restart.
• A valid visible proposal is preserved when a CLI omits or malforms its proposal receipt; strict structured receipts still protect later Auto Discussion phases.

Verification

• Frontend: 306/306 tests passed.
• Desktop: 1338/1338 tests passed.
• Deterministic Eval Harness: 6 cases and 18 results passed.
• Web build, Electron renderer build, desktop pack, release archive generation, archive integrity, signature integrity, packaged application acceptance, and git diff --check passed.
• Packaged acceptance confirmed that an unaddressed Concurrent Responses message targets every group Agent and that finite and No round limit composers use the same surface styling.

Distribution

• Apple silicon macOS only: DMG and ZIP are included with SHA256SUMS.txt.
• The app is ad-hoc signed, not Apple Developer ID signed or notarized. macOS may require Open Anyway on first launch.
• Live Agent behavior still depends on the installed CLI version, authentication, Provider, and declared capabilities. This release does not claim live certification for every supported Agent.
• Meldwork is distributed under the Meldwork Non-Commercial Source License. Commercial use requires prior written permission.

Meldwork V1.0.2

Meldwork V1.0.2

This prerelease upgrades Meldwork from sequential group review into a concurrent, Agent-negotiated collaboration system and completes the related desktop interaction work.

Highlights

• Concurrent Responses freezes one shared task snapshot for every selected Agent, isolates same-batch replies, preserves all selected members, and commits results in stable order.
• Auto Discussion now begins with independent parallel proposals. Agents then challenge one another, negotiate responsibilities, execute distinct work packages, synthesize the result, and independently verify the current candidate.
• Harness V4 adds strict structured receipts, bounded Artifact and Evidence references, delivery watermarks, phase-and-slot operation IDs, serialized Human Gates, and idempotent crash recovery.
• Supported local Agent runtimes now share sanitized streaming answer, plan, status, warning, and tool-lifecycle events without exposing commands, paths, credentials, or private reasoning.
• Unlimited discussions show real state-driven phases for parallel divergence, cross-Agent challenge, coordinated work, synthesis, and verification.
• Group and direct chats now include an animated return-to-bottom control. Message copy, version, regenerate, and delete controls were also corrected for narrow layouts.

Safety And Compatibility

• Existing V1 manual, V2 Auto Discussion, and V3 Task Graph records continue through their original parsers and executors.
• Workspace writes remain opt-in. Concurrent replies, proposals, challenges, negotiated work, and verification are read-only; only the selected synthesis writer can receive write access.
• Renderer-visible state remains limited to validated phases, participants, roles, progress, and Gate data.

Verification

• Frontend: 302/302 tests passed.
• Desktop: 1335/1335 tests passed.
• Deterministic Eval Harness: 6 cases and 18 results passed.
• Web build, Electron renderer build, desktop pack, DMG/ZIP integrity checks, deep code-sign verification, and git diff --check passed.
• The final packaged 360 x 800 acceptance passed all 8 checks for group/direct streaming, return-to-bottom behavior, Trace bounds, message actions, and stop durability.
• Exact source and tag CI passed at commit 053026a736e740ba071335379efdf84fd4e16510.

Current Live Boundaries

• Hermes live streaming and tool lifecycle acceptance passed.
• The latest OpenClaw live run failed closed with LOCAL_AGENT_PROCESS_FAILED. Protocol fixtures and managed-runtime tests pass, but this release does not claim successful live OpenClaw streaming/tool acceptance.
• One Codex direct run completed after exceeding the 300-second observation window; this remains a live latency limitation, not a UI failure.

Distribution

• Apple silicon only: DMG and ZIP are included with SHA256SUMS.txt.
• The app is ad-hoc signed, not Apple Developer ID signed or notarized. macOS may require Open Anyway on first launch.
• Meldwork is distributed under the Meldwork Non-Commercial Source License. Commercial use requires prior written permission.

Meldwork V1.0.1

Meldwork V1.0.1

这是一版面向 Apple 芯片 macOS 的预览发行版，重点改善长时间 Agent 任务、群聊复核和中断恢复体验。

这次更新带来了什么

• 长时间任务更可靠：Agent 只要仍在产生可观察的有效进展，就不会仅因运行时间较长而被误判为超时。
• 无限轮次更严格：选择 unlimited 模式后，参与讨论的 Agents 会主动质疑缺少依据的结论，并在验收阶段严格复核自己和其他 Agents 的工作。
• 群聊过程更容易理解：执行过程、工具活动和结论会保持有序，重要信息可以继续追溯，同时避免把所有过程细节无限塞进上下文。
• 会话操作更清晰：加载状态、会话操作和执行详情更容易扫描，用户可以更快判断当前状态和下一步动作。
• 中断后更容易恢复：已有进展会被保留，任务不会因为一次中断就丢失整个工作脉络。

下载与校验

• Meldwork-0.1.1-arm64.dmg：macOS Apple 芯片安装包
• Meldwork-0.1.1-arm64.zip：macOS Apple 芯片压缩包
• SHA256SUMS.txt：安装包校验值

当前构建未使用 Developer ID 签名，也未经过 Apple 公证。首次启动若被 macOS 拦截，请按 README 中的“仍要打开 / Open Anyway”步骤操作。

验证记录

• 前端测试：263 项通过
• 前端生产构建：通过
• 桌面前端构建：通过
• 桌面测试：993 项中 991 项通过；2 项 CLI 取消测试受本机测试辅助进程环境影响，未涉及本次发布改动。
• DMG 与 ZIP SHA256：已通过本地复核

Meldwork First Version

This release packages the latest Meldwork runtime, local Agent readiness, conversation isolation, and chat UX fixes for Apple silicon macOS.

Highlights

• Isolated direct and group conversation send/run state so one active conversation no longer blocks another.
• Improved local Agent readiness detection from login-shell Provider variables and known native config locations, including OpenRouter/custom Provider cases, without exposing credentials.
• Refined completion UX: per-conversation notifications, right-aligned completion indicators, hover-scoped message actions, and stacked multimodal/document attachments.
• Reorganized desktop main-process modules by domain while preserving narrow renderer IPC boundaries.
• Updated README installation guidance for unsigned macOS preview builds.

Verification

• Frontend tests: 31 files / 259 tests passed.
• Desktop tests: 989 tests passed.
• Deterministic Eval Harness: 6 cases / 18 results.
• Web build and desktop renderer build passed.
• Desktop pack and macOS DMG/ZIP dist passed.
• SHA256SUMS.txt generated and verified locally.

Download

• Meldwork-0.1.0-arm64.dmg for Apple silicon Macs.
• Meldwork-0.1.0-arm64.zip as a portable archive.
• SHA256SUMS.txt for integrity verification.

This preview is distributed without Apple Developer ID signing or notarization. macOS may require opening System Settings → Privacy & Security and choosing Open Anyway after the first blocked launch attempt.