➤ 最近还是不断有新的 Agent Harness 思路和实践在出现。

这两天看到 OpenSquilla，一个开源、能本地托管的 AI Agent。

① 它有智能模型路由——同样的任务，token 成本比 OpenClaw 省 60-80%，跑任务时还有个「老虎机」动画告诉你这一单省了多少钱。

路由省的不只是钱：现在各家模型各有所长、价格也不一样，benchmark 高低跟具体任务上的表现还常常对不上。

与其死磕单一最强模型，不如按任务把活派给最合适的那个——它真正在优化的，是「每块钱 token 换回多少智能」。

② 刚发布 MetaSkill。简单说，它是「Skill 的 Skill」——一份告诉模型怎么检索、筛选、组合原子 Skill 的元 markdown。

因为单个 Skill 只干一件事，搜索归搜索，文档归文档。用户自己当调度员，脑子里一直装着下一步该使唤哪个，Skill 一多就会手忙脚乱。

MetaSkill 接管了这个「项目经理」的角色——由它来安排哪些步骤可以并行、哪些只能串行，让 Agent 学会自己组织技能。

③ 还有 meta-skill-creator：一句话需求，自动合成一个新的 meta-skill，手写 30 分钟的活压到 3 分钟。等于是 Agent 自己造技能。

➤ Why now？三件事撞到了一起：模型已经听得懂复杂的多步骤编排指令；社区 Skill 在爆发式增长，多到必须有个更高的抽象层来筛；大模型在线 trial-and-error 依然太贵，得把优化前置到 Skill 组织层。

这三条线指向同一点——Agent 要解决的问题，正在从「会不会调用工具」变成「会不会组织工具」。

**➤ 更深一层：**Harness 这层会不会只是个过渡概念、迟早被更强的模型吃掉？

我倒不这么看——只要模型还在分化、技能和工具还在指数级变多，「怎么组织、怎么调度」就是个不会消失的问题。

很有趣的方向。

GitHub：

opensquilla/opensquilla

Source: https://github.com/opensquilla/opensquilla

OpenSquilla — Token-Efficient AI Agent

Same budget, more capability, better results.
A microkernel AI agent for your CLI, Web UI, and chat channels.

Overview

OpenSquilla is a token-efficient, microkernel AI agent. A local model router sends each turn to the cheapest model that can handle it, while persistent memory, a layered sandbox, built-in web search, and on-device embeddings round out a single shared turn loop.

Every entry point — Web UI, CLI, and chat channels — runs through that same loop, so tool dispatch, retries, and decision logging behave identically everywhere. A pluggable provider layer speaks to OpenRouter, OpenAI, Anthropic, Ollama, DeepSeek, Gemini, Qwen/DashScope, and 20+ other LLM providers with no change to your code or config schema.

OpenSquilla 0.3.1 is the current release.

For task-oriented product documentation, start with the OpenSquilla Product Guide or the documentation index.

Installation

OpenSquilla runs on Windows, macOS, and Linux. Pick the path that matches your use case.

Windows portable and Quick terminal install give you a prebuilt release — no Git required. The other two — Install from source and Develop from source — build from a Git checkout (git clone + Git LFS).

Release install commands use published GitHub release assets. The Windows portable zip also has a /releases/latest/download/ alias for the current release. Python wheel installs use versioned wheel filenames because installers validate the version embedded in the wheel filename.

Prerequisites

The default recommended profile installs SquillaRouter — OpenSquilla’s on-device model router — and its model assets; OPENSQUILLA_INSTALL_PROFILE=core omits those dependencies. The separate --router disabled onboarding flag keeps the dependencies installed but turns the router off at runtime.

On Windows, SquillaRouter’s bundled ONNX runtime also needs the Visual C++ runtime. The Windows portable launcher and the from-source PowerShell installer install it automatically via winget; the Quick terminal install (uv tool install) path does not — if startup logs a DLL load failed error, install it manually (see Troubleshooting). OpenSquilla keeps running with direct single-model routing until it is installed.

Install links: Git · Git LFS · uv.

Windows portable (no Python)

The fastest path on Windows — the zip ships a bundled CPython runtime, so no separate Python install is required.

1. Download the current portable zip: https://github.com/opensquilla/opensquilla/releases/latest/download/OpenSquilla-windows-x64-portable.zip
2. Extract it to a writable folder such as Downloads or Documents, then right-click Start OpenSquilla.cmd and choose Run as administrator.
3. Complete the first-run setup, then open http://127.0.0.1:18791/control/.

Preview builds are unsigned; administrator launch is the supported path. If SmartScreen appears, choose More info → Run anyway. If Smart App Control or enterprise policy blocks the unsigned app, use Quick terminal install instead.

$env:OPENROUTER_API_KEY="sk-..."
Set-ExecutionPolicy -Scope Process Bypass
.\start.ps1

.\opensquilla.cmd onboard --provider openrouter --api-key-env OPENROUTER_API_KEY

Quick terminal install

The recommended path on Windows, macOS, and Linux. uv installs OpenSquilla into its own isolated environment and manages its own Python — no system Python required. This path installs published releases only; for main, development branches, or local checkouts use Install from source.

1. Install uv — skip if uv --version already works.

Linux / macOS:

curl -LsSf https://astral.sh/uv/install.sh | sh
. "$HOME/.local/bin/env"

Windows PowerShell:

powershell -c "irm https://astral.sh/uv/install.ps1 | iex"
$env:Path = "$env:USERPROFILE\.local\bin;" + $env:Path

2. Install OpenSquilla — the same command on every platform.

uv tool install --python 3.12 "opensquilla[recommended] @ https://github.com/opensquilla/opensquilla/releases/download/v0.3.1/opensquilla-0.3.1-py3-none-any.whl"

This installs the OpenSquilla wheel from the release URL, then lets uv download the dependencies declared by the selected extras. The default recommended extra includes SquillaRouter runtime dependencies such as ONNX Runtime, LightGBM, NumPy, and tokenizers, so a first install needs network access unless those wheels are already cached.

3. Configure and run.

opensquilla onboard
opensquilla gateway run

If opensquilla is not found right after a fresh uv install, open a new terminal, or re-run the PATH line from step 1.

For a fully pinned install, use the versioned wheel URL: https://github.com/opensquilla/opensquilla/releases/download/v0.3.1/opensquilla-0.3.1-py3-none-any.whl.

Install from source

Use this path to run OpenSquilla from a checkout without editing it. The clone is only the package source for the installer; after install, use the opensquilla command — do not run uv run. Choose Develop from source instead if you intend to modify the code.

1. Clone with LFS assets

   git lfs install
   git clone https://github.com/opensquilla/opensquilla.git
   cd opensquilla
   git lfs pull --include="src/opensquilla/squilla_router/models/**"
   

2. Run the installer

   macOS / Linux

   bash scripts/install_source.sh
   

   Windows PowerShell

   powershell -ExecutionPolicy Bypass -File ./scripts/install_source.ps1
   

   The script installs .[recommended] (SquillaRouter + memory + local models) into a dedicated user environment via uv tool install, falling back to python -m pip install --user when uv is unavailable. Open a new terminal if opensquilla is not on PATH after install.

3. (optional) Install advanced extras. Most channels — Feishu, Telegram, DingTalk, QQ, WeCom, Slack, and Discord — work from the base install. The opt-in extras are:

   • matrix — Matrix channel (pulls in matrix-nio)
   • matrix-e2e — Matrix channel with end-to-end encryption (requires libolm)
   • document-extras — PDF generation via WeasyPrint

   OPENSQUILLA_INSTALL_EXTRAS=matrix bash scripts/install_source.sh        # macOS / Linux
   

   powershell -ExecutionPolicy Bypass -File ./scripts/install_source.ps1 -Extras matrix   # Windows
   

4. Configure and run — see Configuration.

winget install --id Git.Git -e
winget install --id GitHub.GitLFS -e
powershell -ExecutionPolicy Bypass -c "irm https://astral.sh/uv/install.ps1 | iex"
git lfs install

brew install git git-lfs uv
git lfs install

sudo apt update && sudo apt install -y git git-lfs
curl -LsSf https://astral.sh/uv/install.sh | sh
git lfs install

OPENSQUILLA_INSTALL_PROFILE=core   bash scripts/install_source.sh   # minimal runtime, no SquillaRouter
OPENSQUILLA_INSTALL_DRY_RUN=1      bash scripts/install_source.sh   # print the plan only

Develop from source

Use this path when you are working on OpenSquilla’s source code: making changes, running tests, or debugging behavior against this checkout. It is not the normal install path. Unlike Install from source, this path requires uv: uv sync creates a repository-local .venv, and uv run executes commands against the files in this checkout.

uv sync --extra recommended --extra dev
uv run opensquilla --help

The recommended extra includes SquillaRouter for development too; the dev extra installs the test, lint, and typecheck tools. Install additional extras into the same environment you run:

uv sync --extra recommended --extra dev --extra matrix
uv run opensquilla channels status matrix --json

In this mode, prefix every opensquilla command in Configuration with uv run. Do not debug a development checkout through a user-local opensquilla command — that command runs in a different Python environment.

Configuration

First-run setup

opensquilla onboard is the interactive first-run wizard. It writes the active config file and keeps provider secrets in environment variables when you pass --api-key-env. The router defaults to recommended (SquillaRouter on supported providers); pass --router disabled for direct single-model routing.

opensquilla onboard                # full interactive wizard
opensquilla onboard --if-needed    # idempotent: safe for scripts and re-installs
opensquilla onboard --minimal      # provider only; skip channels and search
opensquilla onboard status         # inspect every setup section without writing

In SSH, CI, or any environment without a TTY, use the non-interactive form — keep the secret in the environment and pass its name, not its value:

Linux / macOS

export OPENROUTER_API_KEY="sk-..."
opensquilla onboard --provider openrouter --api-key-env OPENROUTER_API_KEY

Windows PowerShell

$env:OPENROUTER_API_KEY="sk-..."
opensquilla onboard --provider openrouter --api-key-env OPENROUTER_API_KEY

OpenRouter is only an example — substitute any supported provider and its API-key variable.

Re-configure one section later without redoing the whole wizard (these examples assume the relevant API key is already in the environment):

opensquilla configure provider --provider openai --model gpt-4o --api-key-env OPENAI_API_KEY
opensquilla configure router --router recommended
opensquilla configure search   --search-provider brave --api-key-env BRAVE_SEARCH_API_KEY
opensquilla configure channels

Sections: provider, router, channels, search, image-generation, memory-embedding. The Web UI exposes the same catalog and status model at /control/setup: Provider and Router are the fast path, while Channels, Search, Image generation, and Memory embedding sit in the Capability Center and can be configured later. Empty channels are treated as an opt-out, not a failed setup.

Config load order: OPENSQUILLA_GATEWAY_CONFIG_PATH → ./opensquilla.toml → ~/.opensquilla/config.toml → built-in defaults. Environment values for individual secrets always win over file values.

Migrate from OpenClaw or Hermes Agent

If you already have state under ~/.openclaw or ~/.hermes, run a dry run first to inspect the migration report, then apply it explicitly:

opensquilla migrate openclaw --json
opensquilla migrate openclaw --apply

opensquilla migrate hermes --json
opensquilla migrate hermes --apply

Use opensquilla migrate --source openclaw,hermes --apply to import both default homes. Add --migrate-secrets only after reviewing the dry-run report. See MIGRATION.md for custom paths and conflict handling.

Run

opensquilla gateway run                # foreground, 127.0.0.1:18791
opensquilla gateway start --json       # background + health wait
opensquilla chat                       # interactive REPL
opensquilla agent -m "your prompt"     # one-shot, automation-friendly

Open the Web UI at http://127.0.0.1:18791/control/. The Health view shows whether OpenSquilla is ready, what is not ready, and the next recovery steps. From the CLI, run:

opensquilla doctor
opensquilla doctor --json
opensquilla doctor --config ./opensquilla.toml --json

/health and /healthz are lightweight liveness endpoints for process checks. opensquilla doctor and the Web UI Health view are the readiness surfaces for provider config, memory, logs, search, channels, sandbox posture, router, image generation, and recovery guidance. Press Ctrl+C to stop a foreground gateway.

Other command groups include sessions, skills, memory, migrate, cron, channels, providers, models, and cost. Run opensquilla --help or opensquilla <group> --help for details.

opensquilla gateway restart
opensquilla channels status <name> --json

opensquilla gateway run --listen 0.0.0.0 --port 18791

docker build -t opensquilla:local .

Provider tiers, sandbox tuning, image generation, and concurrency settings live in opensquilla.toml.example.

What’s New in 0.3.1

OpenSquilla 0.3.1 is a maintenance release for the 0.3 line. It updates the stable install metadata and brings selected channel, chat, provider, and workflow fixes from the integration branch onto the stable release line:

• Channel setup and replies — Slack Socket Mode, app mentions, signing secrets, and threaded replies preserve the channel context needed for setup and replies.
• Media and voice workflow handoffs — short-drama/video helper workflows remain bundled, generated media flows have clearer review pauses, and voice/audio handoffs are usable end to end.
• Chat formatting — user message bubbles preserve multiline text and read like authored messages instead of compressed UI labels.
• Provider request hardening — malformed tool-call history is kept away from providers before it becomes invalid request state.
• Install and release checks — installer URLs, release metadata, version consistency tests, and CI impact gates are updated for 0.3.1.

Full notes: CHANGELOG.md · docs/releases/0.3.1.md.

What’s New in 0.2.1

OpenSquilla 0.2.1 is a maintenance release focused on release-package startup and long-running agent reliability:

• Windows portable startup — the portable launcher better detects and bootstraps the Visual C++ runtime needed by the bundled ONNX router.
• Long-running agent turns — tool-heavy WebUI sessions recover more cleanly from oversized tool results, malformed tool calls, artifact delivery handoffs, and degraded final responses.
• Cleaner WebUI output — generated artifact markers are kept out of normal chat replay while delivered files remain visible.
• Memory recall scoring — local and OpenAI-compatible embedding vectors are normalized before semantic search, and strong keyword matches remain usable when vector scores are low.

Full notes: CHANGELOG.md · release notes.

What’s New in 0.2.0

This release expands OpenSquilla across migration, CLI chat, channels, scheduling, and long-running tool work:

• Migration path from existing agent homes — opensquilla migrate previews and applies imports from existing OpenClaw/Hermes homes, including memory, persona files, skills, MCP/channel config, conflict handling, and migration reports.
• Usable chat CLI — opensquilla chat now has a persistent terminal UI, streaming output, queued input, slash-mode discovery, tool/status strips, and more deterministic live prompt behavior.
• Cross-surface cron automation — cron jobs now cover structured schedules, timezone-aware exact/every/cron runs, channel or webhook delivery, failure destinations, manual runs, and WebUI/CLI/RPC parity.
• Better Feishu and Discord channels — channel adapters expose clearer capability metadata, safer DM/group handling, native file and artifact paths, and improved attachment/thread behavior while privileged actions stay scoped.
• Sturdier long-running turns — failed turns are kept out of provider replay, malformed tool calls are handled more safely, and approval-gated retries wait for operator decisions.
• Smarter context and tool budgeting — provider-budget compaction, prompt cache preservation, bounded tool results, and side-effect-aware concurrency make large tool-heavy sessions more predictable.
• Web UI and release polish — recency ordering, table layout, mobile controls, duplicate notifications, setup forms, release URLs, and install paths are tightened for 0.2.0.

Full notes: CHANGELOG.md · release notes.

Key Features

MetaSkill docs: docs/features/meta-skills.md, docs/features/meta-skill-user-guide.md, and docs/authoring/meta-skills.md.

Benchmark Results

PinchBench 1.2.1 average results across 25 tasks:

Score is the mean across the 25 tasks; token counts and cost are totals for the full run.

Troubleshooting

opensquilla onboard --provider openrouter --api-key-env OPENROUTER_API_KEY --router recommended
opensquilla gateway restart

Credits

OpenSquilla is inspired by OpenClaw. Bundled third-party content is attributed in THIRD_PARTY_NOTICES.md.

Community contributors are acknowledged in CONTRIBUTORS.md, including release-specific attribution notes for squash-merged or replayed work.

Contributing

Contributions of every kind are welcome — bug reports, feature ideas, documentation, new provider or channel adapters, skills, and core runtime work. See CONTRIBUTING.md, then open an issue or pull request on GitHub.

Code of Conduct · Security · Support · License (Apache-2.0)