v0.1.9 — Skill Workflows and Provider Key Updates

v0.1.9 — Skill Workflows and Provider Key Updates

OpenHarness v0.1.9 is a small follow-up release after v0.1.8 focused on making skills easier to create and invoke, plus fixing provider API key updates.

Highlights

• Bundled skill creator

  • Added a bundled skill-creator skill for creating, improving, and verifying OpenHarness/ohmo skills.
  • This makes repeatable workflows easier to capture as first-class skills.

• User-invocable skill slash commands

  • Skills marked as user-invocable can now be triggered directly with slash commands.
  • Slash-invoked skills support user arguments and can request a model override through skill metadata.

• Provider API key update fix

  • oh setup now lets users update the API key for an already-configured API-key provider profile.
  • oh provider edit <profile> --api-key <key> can replace a saved profile key directly.
  • oh provider add ... --api-key <key> can store a key while creating a provider profile.

Fixes

• Fixed issue #238, where users could change a configured provider model but had no supported path to update the saved API key.

Install

pip install --upgrade openharness-ai==0.1.9

Contributors

This release is primarily a maintainer follow-up release. Thanks to the users and contributors who reported and validated the provider key update workflow, especially the reporter of #238.

v0.1.8 — Providers, ohmo Feishu Groups, and Safer Remotes

v0.1.8 — Providers, ohmo Feishu Groups, and Safer Remotes

OpenHarness v0.1.8 is a stabilization and provider-expansion release. It adds more first-class provider workflows, improves ohmo's Feishu group experience, hardens remote-channel security, and fixes several Windows/MCP reliability issues.

Highlights

• New provider workflows

  • Added NVIDIA NIM as a built-in OpenAI-compatible provider using NVIDIA_API_KEY.
  • Added ModelScope Inference API support.
  • Added Qwen (DashScope), MiniMax, and Gemini provider profiles.
  • Improved OpenAI-compatible API behavior, including explicit bearer authorization headers and <think> block filtering for compatible streaming responses.

• ohmo Feishu group support

  • Added Feishu managed group creation flow for ohmo.
  • Added gateway-scoped provider/model commands for chat-based operation.
  • Improved group routing and mention policy so ohmo responds in shared Feishu groups only when explicitly addressed.
  • Hardened Feishu attachment filename handling.

• Security and remote-channel hardening

  • Kept sensitive config/auth/provider/model/ship commands local-only by default in remote channels.
  • Kept bridge commands local-only by default.
  • Added coverage for bridge spawn blocking and remote gateway security behavior.
  • Rejected path traversal names during plugin uninstall.

• Windows and shell reliability

  • Fixed Windows agent/subagent spawning by direct-executing teammate argv instead of shell-wrapping Python paths.
  • Windows shell resolution now skips discovered bash.exe binaries that cannot actually run commands, falling back to PowerShell/cmd.
  • Improved Windows gateway process lifecycle handling.
  • Avoided shell execution when opening browsers on Windows.

• MCP, tools, and stability

  • MCP startup now isolates failed servers instead of aborting the whole OpenHarness startup.
  • Fixed subprocess stderr pipe deadlocks in grep/glob/bash/session runner paths.
  • Bounded large tool results in conversation history and improved compaction under large/vision contexts.
  • Improved skill frontmatter parsing with YAML safe_load for bundled and user skills.

• TUI and UX fixes

  • Restored raw DEL backspace handling for macOS Terminal-style environments.
  • Added better slash-command completion, markdown table rendering, spinner behavior, and escape interruption.
  • Added image-to-text fallback support for text-only models and --vision-model override.

External contributors

Thanks to the external contributors whose PRs are included in this release:

• @Litianhui888 — MCP startup isolation (#237)
• @Hinotoi-agent — remote command security hardening (#232, #209, #208, #198, #197)
• @nsxdavid — Windows agent spawn fix (#231)
• @voidborne-d — bundled skill frontmatter parsing (#229)
• @Mcy0618 — image-to-text fallback and ModelScope support (#227, #224)
• @WANG-Guangxin — invalid regex fallback fix (#219)
• @glitch-ux — Windows browser auth safety, async coordinator draining, autopilot shell safety, hook events (#217, #200, #188, #170)
• @Escapingbug — Windows gateway lifecycle and Telegram proxy config fixes (#193, #192)
• @ZevGit — Qwen provider profile (#207)
• @yl-jiang — subprocess stderr deadlock fixes and OpenAI-compatible think-block filtering (#205, #174)
• @he-yufeng — TUI slash-command completion polish (#185)
• @powAu3 — raw DEL backspace fix (#182)
• @flobo3 — shell subprocess stdin default fix (#179)

Install

pip install --upgrade openharness-ai==0.1.8

Or run the installer from the repository docs.

v0.1.7 — TUI Polish & Safer Install

Highlights for 0.1.7

• Safer install flow: scripts/install.sh now links oh, ohmo, and openharness into ~/.local/bin instead of prepending the virtualenv bin directory to PATH, which avoids clobbering Conda-managed shells.
• Better React TUI input: Shift+Enter now inserts a newline while plain Enter still submits.
• Quieter busy-state animation: the extra pseudo-animation line was removed, and Windows terminals now use conservative ASCII spinner frames to reduce flashing.
• PyPI package published: openharness-ai==0.1.7.

What's Changed

• feat(tui): pair tool-call and tool-result rows for cleaner transcript by @yl-jiang in #111
• fix(auth): cache keyring probe and rename misleading encrypt/decrypt by @glitch-ux in #106
• feat(providers): add built-in Google Gemini provider profile by @glitch-ux in #105
• fix: regenerate system prompt on checkpoint restore by @JiangweiYe76 in #104
• fix(tui): keep busy spinner active throughout agent turn by @yl-jiang in #116
• fix(config): strip ANSI escape sequences from model names by @jiakeboge in #114
• fix(ui): show effective runtime model in header by @siaochuan in #95
• fix(grep): handle long rg lines without crashing by @tjb-tech in #112
• feat(personalization): auto-extract local environment rules from session history by @siaochuan in #65
• fix(grep): raise asyncio stream limit and catch ValueError on long lines by @yl-jiang in #110
• fix(skills): use yaml.safe_load for SKILL.md frontmatter parsing by @fengjie926 in #96
• [security] Harden path rules and web fetch network guards by @13ernkastel in #92
• bugfix: fix Claude subscription auth lookup on macOS by @hackereee in #123
• feat(sandbox): add Docker as an alternative sandbox backend by @glitch-ux in #121
• fix(persistence): atomic writes and locks for shared state by @glitch-ux in #128
• fix: harden gateway slash command security by @Hinotoi-agent in #127
• fix(feishu): reply in group threads instead of creating new topics by @hehe1111 in #125
• fix(tui): write trailing newline on exit so shell prompt starts on a fresh line by @yl-jiang in #133
• [Windows] Add PowerShell installer for native Windows support - Add … by @benben951 in #118
• fix(tools): todo_write updates existing items in-place instead of duplicating by @yl-jiang in #135
• fix(engine): keep parallel tool turns alive when one tool raises by @glitch-ux in #138
• feat(tui): add codex output style to mitigate streaming flicker by @siaochuan in #141
• fix(installer): detect installed launcher instead of hard-coding openh by @glitch-ux in #145
• [security] fix(ohmo): secure default remote channel allowlists by @Hinotoi-agent in #147
• fix(runtime): resolve profile base_url and support openai_compat format by @yay2008 in #146
• [codex] fix agent spawn overrides and task cleanup by @yu2001-s in #148
• fix(settings): Add Ollama Support and prevent env overrides from clobbering provider profile… by @forrestzhang in #139
• feat: add MiniMax as a first-class provider by @octo-patch in #140
• fix(agent): Explore and claude-code-guide no longer hard-code model=haiku by @yl-jiang in #154
• [security] fix(ohmo): isolate shared-chat sessions by sender by @Hinotoi-agent in #159
• [security] fix(plugin): address code execution via untrusted plugin activation by @Hinotoi-agent in #156
• [security] fix(personalization): stop replaying exported env var values by @shaun0927 in #151
• [security] fix(sandbox): fail closed for unsupported Docker domain policies by @shaun0927 in #152

New Contributors

• @JiangweiYe76 made their first contribution in #104
• @jiakeboge made their first contribution in #114
• @fengjie926 made their first contribution in #96
• @hackereee made their first contribution in #123
• @Hinotoi-agent made their first contribution in #127
• @hehe1111 made their first contribution in #125
• @benben951 made their first contribution in #118
• @yay2008 made their first contribution in #146
• @yu2001-s made their first contribution in #148
• @octo-patch made their first contribution in #140
• @shaun0927 made their first contribution in #151

Full Changelog: v0.1.6...v0.1.7

v0.1.6 — ohmo Personal Agent & Auto-Compaction

Highlights

ohmo — a personal AI agent built on OpenHarness. Chat with ohmo in Feishu / Slack / Telegram / Discord, and it forks branches, writes code, runs tests, and opens PRs on its own. ohmo runs on your existing Claude Code or Codex subscription — no extra API key needed.

pip install openharness-ai
oh setup
oh

ohmo init && ohmo config && ohmo gateway start

---

What's New since v0.1.0

🤖 Swarm & Multi-Agent

• Full coordinator orchestration system with agent definition loading
• InProcessBackend with contextvars isolation, subprocess backend for pollable agents
• WorktreeManager with git worktree isolation per teammate
• TeamLifecycleManager with persistent team file management
• Mailbox communication system with file-lock layer extracted to standalone module
• Permission sync protocol for leader-worker bridging
• Subprocess teammates run in headless worker mode

🧠 Auto-Compaction

• Context auto-compression for long-running sessions
• Task state and channel logs preserved across compaction
• Tested: ohmo ran 4 days straight without manual compact or clear

📡 ohmo Personal Agent

• ohmo init / ohmo config / ohmo gateway start — 3 commands to go live
• 10 IM channels: Telegram, Slack, Discord, Feishu, DingTalk, WeChat, QQ, Matrix, WhatsApp, Email
• Channel slash commands, file attachments, multimodal gateway messages
• Runs on Claude Code subscription or Codex subscription

🔌 MCP

• HTTP transport support
• Auto-reconnect on disconnected servers
• JSON Schema type inference for tool inputs
• Tool-only server compatibility

🌙 Multi-Provider

• Native Moonshot / Kimi support with thinking model reasoning_content handling
• GitHub Copilot provider with OAuth device flow
• OpenAI-compatible API client (--api-format openai)
• Unified oh setup wizard for all providers
• Profile-scoped credentials — different endpoints no longer share one global key

🎨 TUI

• 5 switchable themes (oh --theme cyberpunk)
• TodoPanel, SwarmPanel, PlanMode indicator, enhanced AskUser modal
• Assistant messages render full Markdown in React TUI
• Permission modals serialized to prevent input swallowing
• Rendering throttle for smoother streaming

🛡️ Security & Stability

• Built-in sensitive path protection in PermissionChecker
• Shell injection fix in command hook $ARGUMENTS substitution
• web_fetch URL validation hardened
• EIO crash recovery in Ink TUI
• Windows: UTF-8 protocol, cmd flash fix, git prompt deadlock fix, WSL dialog fix
• Network error notification with real error messages
• glob/grep backed by ripgrep (10x faster)

📦 Release

• pip install openharness-ai on PyPI
• Frontend bundled in wheel
• One-click installer: curl -fsSL .../install.sh | bash
• oh --version / --resume / --continue flags
• Cron scheduler daemon

---

182 commits · 242 files changed · 41,000+ lines added

https://github.com/HKUDS/OpenHarness