root/fusion
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

fix(plugins): re-export probe symbols + declare plugin deps in dashboard

Browse Source 
- Hermes / OpenClaw plugin index.ts now re-export `probeHermesBinary` /
  `probeOpenClawBinary` and their status types so the dashboard's
  `runtime-provider-probes.ts` façade can import them via the public
  package entry instead of deep paths.
- Dashboard `package.json` adds `@fusion-plugin-examples/hermes-runtime`,
  `…/openclaw-runtime`, `…/paperclip-runtime` as workspace deps so
  pnpm symlinks them into `packages/dashboard/node_modules/`. Without
  these, the new probe imports failed with "Cannot find module" during
  `pnpm typecheck`.

This clears 6 of the 9 outstanding typecheck errors. The remaining 3 are
in the in-flight Hermes plugin rewrite (runtime-adapter still imports
from a deleted `./pi-module.js`; the new `index.ts` calls a factory
with the wrong arg type) and should be resolved by the same change set
that landed the rewrite.

Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This commit is contained in:
Fusion
2026-04-27 20:22:18 -07:00
committed by gsxdsm
parent 5159cc2a68
commit 6e65786bc5
82 changed files with 8481 additions and 2833 deletions
Download Patch File Download Diff File Expand all files Collapse all files

131
plugins/fusion-plugin-openclaw-runtime/README.md
View File

@@ -1,79 +1,77 @@
# OpenClaw Runtime Plugin
`fusion-plugin-openclaw-runtime` provides the `openclaw` runtime hint for Fusion agents by calling a **locally running OpenClaw gateway** over HTTP.
Drives the local **`openclaw` CLI** ([openclaw/openclaw](https://github.com/openclaw/openclaw)) as a subprocess. By default it runs `openclaw agent --local` (embedded mode, no daemon required); you can opt into the WebSocket gateway with `useGateway: true`.
Unlike the default runtime, this plugin does **not** delegate to Fusion's internal pi runtime. It talks directly to OpenClaw's OpenAI-compatible endpoint.
## What it does
For each `promptWithFallback(session, prompt)`:
1. Spawns `openclaw --no-color agent --local --json --session-id <uuid> --message <prompt>` (plus `--agent`, `--model`, `--thinking`, `--timeout` if configured).
2. Reads the single JSON document on stdout (matching `OpenClawAgentJson`):
- Concatenates `payloads[]` where `!isError && !isReasoning` → `session.callbacks.onText(...)`.
- Joins `payloads[].isReasoning === true` → `session.callbacks.onThinking(...)`.
- Surfaces tool-level errors (`payloads[].isError === true`) as a logger warning.
- Stores `meta.agentMeta.usage` on the session for token accounting.
3. Reuses the same UUID across every prompt for the session so OpenClaw resumes the same agent conversation server-side.
The previous HTTP `/v1/chat/completions` integration has been removed — that endpoint required a separate gateway daemon and was an OpenAI-compat shim. The CLI surface is the canonical OpenClaw API.
## Prerequisites
1. Install OpenClaw globally:
```bash
npm i -g openclaw
npm install -g openclaw
```
2. Start your OpenClaw gateway with chat-completions endpoint enabled.
3. Ensure the gateway is reachable from Fusion (default: `http://127.0.0.1:18789`).
Verify with `openclaw --version` (expect `OpenClaw 2026.x.y`).
## Installation
If you want gateway mode (`useGateway: true`), also start `openclaw gateway run` separately.
### Option 1: Copy to plugins directory
> **First-run note:** the very first `openclaw agent` invocation lazy-installs runtime deps and can take 30–60s. Subsequent calls are fast.
```bash
cp -r fusion-plugin-openclaw-runtime ~/.fusion/plugins/
## Settings
| Key | Env var | Default | Notes |
|---|---|---|---|
| `binaryPath` | `OPENCLAW_BIN` | `openclaw` | Path to the `openclaw` binary. |
| `agentId` | `OPENCLAW_AGENT_ID` | `main` | Maps to `--agent <id>`. List with `openclaw agents list`. |
| `model` | `OPENCLAW_MODEL` | (OpenClaw default) | Maps to `--model <provider/model>`, e.g. `anthropic/claude-haiku-4-5`. |
| `thinking` | `OPENCLAW_THINKING` | `off` | One of `off | minimal | low | medium | high | xhigh | adaptive | max`. |
| `cliTimeoutSec` | `OPENCLAW_TIMEOUT_SEC` | `0` | OpenClaw-side timeout (0 = no limit). |
| `cliTimeoutMs` | `OPENCLAW_CLI_TIMEOUT_MS` | `300000` | Hard kill on the Fusion side. |
| `useGateway` | `OPENCLAW_USE_GATEWAY` | `false` | When `true`, omit `--local`; the CLI tries the WS gateway and falls back to embedded after ~2 s. |
Settings precedence: plugin settings → env var → default.
## Limitations
- **No per-token streaming.** `--json` emits a single JSON document at process exit. `onText` is called exactly once.
- **Default ignores the gateway.** With `useGateway: false` (default) we always pass `--local`, skipping the WebSocket connect attempt entirely. Most users want this.
- **AbortSignal sends SIGTERM.** If the CLI ignores it (e.g. during a long model download), the hard-kill timer (`cliTimeoutMs`) eventually fires.
## Public API
```ts
import {
OpenClawRuntimeAdapter,
resolveCliConfig,
buildOpenClawArgs,
createCliSession,
promptCli,
describeCliModel,
extractStderrError,
probeOpenClawBinary,
type CliConfig,
type GatewaySession,
type OpenClawAgentJson,
type OpenClawBinaryStatus,
} from "@fusion-plugin-examples/openclaw-runtime";
```
### Option 2: Install via CLI
`probeOpenClawBinary({ binaryPath?, timeoutMs? })` runs `openclaw --version` and returns `{ available, version, binaryPath, reason, probeDurationMs }` — used by the dashboard's "Runtimes → OpenClaw" settings card.
```bash
fn plugin install ./plugins/fusion-plugin-openclaw-runtime
```
## Agent configuration
## Runtime Metadata
- **Plugin ID:** `fusion-plugin-openclaw-runtime`
- **Package name:** `@fusion-plugin-examples/openclaw-runtime`
- **Runtime ID:** `openclaw`
- **Runtime name:** `OpenClaw Runtime`
- **Version:** `0.1.0`
## Plugin Settings
Configure via plugin settings (`ctx.settings`) or environment variables.
| Setting key | Env fallback | Default |
| --- | --- | --- |
| `gatewayUrl` | `OPENCLAW_GATEWAY_URL` | `http://127.0.0.1:18789` |
| `gatewayToken` | `OPENCLAW_GATEWAY_TOKEN` | _unset_ |
| `agentId` | `OPENCLAW_AGENT_ID` | `main` |
Settings take precedence over environment variables.
## How Execution Works
For each prompt, the runtime sends a streaming request to:
- `POST /v1/chat/completions`
- `Content-Type: application/json`
- `Authorization: Bearer <gatewayToken>` (when configured)
- `x-openclaw-agent-id: <agentId>`
Request payload includes:
- `model: "openclaw:<agentId>"`
- `stream: true`
- `messages: [...]`
- `user: <stable-session-id>` (so repeated turns share a stable gateway session)
Streaming uses SSE (`data: ...` + `[DONE]`), with callbacks wired for:
- text deltas (`choices[0].delta.content`)
- reasoning deltas (`choices[0].delta.reasoning_content`)
- tool call lifecycle (`choices[0].delta.tool_calls`)
## Agent Configuration
Configure an agent to target OpenClaw via `runtimeConfig.runtimeHint`:
To create a Fusion agent backed by OpenClaw, set `runtimeConfig.runtimeHint`:
```json
{
@@ -85,14 +83,17 @@ Configure an agent to target OpenClaw via `runtimeConfig.runtimeHint`:
}
```
## Notes
Runtime selection happens in the dashboard's **New Agent → Plugin Runtime → OpenClaw**.
- This plugin no longer depends on `@fusion/engine`.
- Session cleanup is a no-op client-side; OpenClaw manages gateway sessions.
## Metadata
## Local Development
- **Plugin ID:** `fusion-plugin-openclaw-runtime`
- **Runtime ID:** `openclaw`
- **Package:** `@fusion-plugin-examples/openclaw-runtime`
## Development
```bash
pnpm --filter @fusion-plugin-examples/openclaw-runtime test
pnpm --filter @fusion-plugin-examples/openclaw-runtime test # 44 tests
pnpm --filter @fusion-plugin-examples/openclaw-runtime build
```