openrune 0.2.1 → 0.3.0

package/README.md

@@ -44,10 +44,6 @@ Building a Claude Code harness usually means wiring up process management, I/O p
 
 **Scriptable** — Chain agents, set up triggers, or call agents from your own code via the Node.js API. One file format, multiple ways to use it.
 
-<p align="center">
-  <img src="demo.gif" width="100%" alt="Rune demo" />
-</p>
-
 ---
 
 ## Quick Start
@@ -63,31 +59,24 @@ Building a Claude Code harness usually means wiring up process management, I/O p
 npm install -g openrune
 ```
 
-### 2. Create your first agent
+### 2. Create an agent
 
 ```bash
 cd ~/my-project
-rune new myagent
+rune new reviewer --role "Code reviewer, security focused"
 ```
 
-Or right-click any folder in Finder → Quick Actions → **New Rune**
-
-<p align="center">
-  <img src="Screenshot.png" width="500" alt="Right-click to create a Rune agent" />
-</p>
-
-### 3. Use it
-
-**Desktop UI** — Double-click the `.rune` file, or:
+### 3. Run it
 
 ```bash
-rune open myagent.rune
-```
+# Headless — run from CLI, get results in your terminal
+rune run reviewer.rune "Review the latest commit"
 
-**Headless** — Run from the CLI without a GUI:
+# Pipe input from other commands
+git diff | rune run reviewer.rune "Review this diff"
 
-```bash
-rune run myagent.rune "Explain this project's architecture"
+# Desktop UI — open the chat interface
+rune open reviewer.rune
 ```
 
 ---
@@ -120,71 +109,7 @@ rune watch ops.rune --on cron --interval 10m --prompt "Check if the API is healt
 
 ---
 
-## Features
-
-### Harness
-
-- **Headless execution** — `rune run` lets you run agents from the CLI, scripts, or CI/CD. No GUI needed.
-- **Agent chaining** — `rune pipe` connects agents in sequence. Each agent's output feeds into the next.
-- **Automated triggers** — `rune watch` runs agents on file changes, git commits, or a cron schedule.
-- **Node.js API** — `require('openrune')` to use agents programmatically in your own code.
-- **Persistent context** — Role, memory, and chat history are saved in the `.rune` file across sessions.
-
-### Desktop UI
-
-- **Chat interface** — Markdown rendering, file attachment, stream cancellation.
-- **Real-time activity** — See every tool call, result, and permission request as it happens via [Claude Code hooks](https://docs.anthropic.com/en/docs/claude-code/hooks).
-- **Built-in terminal** — Raw Claude Code output and your own commands, side by side.
-
-### Agent Roles
-
-```bash
-# General assistant
-rune new assistant
-
-# Specialized agents
-rune new designer --role "UI/UX design expert"
-rune new backend --role "Backend developer, Node.js specialist"
-rune new reviewer --role "Code reviewer, focused on security and performance"
-```
-
-### The `.rune` File
-
-A `.rune` file is just JSON:
-
-```json
-{
-  "name": "myagent",
-  "role": "General assistant",
-  "icon": "bot",
-  "createdAt": "2025-01-01T00:00:00Z",
-  "history": []
-}
-```
-
-Edit the `role` field anytime to change the agent's behavior.
-
----
-
-## CLI Commands
-
-| Command | Description |
-|---------|-------------|
-| `rune install` | Build app, register file association, install Quick Action |
-| `rune new <name>` | Create a `.rune` file in the current directory |
-| `rune new <name> --role "..."` | Create with a custom role |
-| `rune open <file.rune>` | Open a `.rune` file (desktop GUI) |
-| `rune run <file.rune> "prompt"` | Run agent headlessly (no GUI) |
-| `rune pipe <a.rune> <b.rune> "prompt"` | Chain agents in a pipeline |
-| `rune watch <file.rune> --on <event>` | Set up automated triggers |
-| `rune list` | List `.rune` files in the current directory |
-| `rune uninstall` | Remove Rune integration (keeps your `.rune` files) |
-
----
-
-## Harness Mode
-
-Rune isn't just a desktop app — it's a full agent harness. Use it from scripts, CI/CD, or your own tools.
+## Harness
 
 ### Headless execution
 
@@ -245,6 +170,55 @@ const { finalOutput } = await rune.pipe(
 
 ---
 
+## Desktop UI
+
+Rune also includes a desktop app for interactive use. Double-click a `.rune` file or run `rune open`.
+
+- **Chat interface** — Markdown rendering, file attachment, stream cancellation.
+- **Real-time activity** — See every tool call, result, and permission request as it happens via [Claude Code hooks](https://docs.anthropic.com/en/docs/claude-code/hooks).
+- **Built-in terminal** — Raw Claude Code output and your own commands, side by side.
+- **Right-click to create** — macOS Quick Action lets you create agents from Finder.
+
+<p align="center">
+  <img src="demo.gif" width="100%" alt="Rune demo" />
+</p>
+
+---
+
+## The `.rune` File
+
+A `.rune` file is just JSON:
+
+```json
+{
+  "name": "reviewer",
+  "role": "Code reviewer, security focused",
+  "createdAt": "2025-01-01T00:00:00Z",
+  "history": [],
+  "memory": []
+}
+```
+
+Edit the `role` field anytime to change the agent's behavior. History and memory persist across sessions automatically.
+
+---
+
+## CLI Commands
+
+| Command | Description |
+|---------|-------------|
+| `rune new <name>` | Create a `.rune` file in the current directory |
+| `rune new <name> --role "..."` | Create with a custom role |
+| `rune run <file.rune> "prompt"` | Run agent headlessly (no GUI) |
+| `rune pipe <a.rune> <b.rune> "prompt"` | Chain agents in a pipeline |
+| `rune watch <file.rune> --on <event>` | Set up automated triggers |
+| `rune open <file.rune>` | Open a `.rune` file (desktop GUI) |
+| `rune list` | List `.rune` files in the current directory |
+| `rune install` | Build app, register file association, install Quick Action |
+| `rune uninstall` | Remove Rune integration (keeps your `.rune` files) |
+
+---
+
 ## Architecture
 
 ```
@@ -273,8 +247,8 @@ const { finalOutput } = await rune.pipe(
 
 **Two modes of operation:**
 
-1. **Desktop UI** — Chat input → MCP channel → Claude Code, with hooks for real-time activity monitoring.
-2. **Harness** — Direct CLI execution via `claude -p`. Agents run headlessly with context from the `.rune` file.
+1. **Harness** — Direct CLI execution via `claude -p`. Agents run headlessly with context from the `.rune` file.
+2. **Desktop UI** — Chat input → MCP channel → Claude Code, with hooks for real-time activity monitoring.
 
 ---
 
@@ -319,12 +293,6 @@ Rune/
       lib/                 # Utilities
 ```
 
-### Hooks Configuration
-
-Rune automatically sets up Claude Code hooks in `~/.claude/settings.json`. The hooks only fire when `RUNE_CHANNEL_PORT` is set, so they don't affect standalone Claude Code usage.
-
-Captured events: `PreToolUse`, `PostToolUse`, `PermissionRequest`, `UserPromptSubmit`, `Stop`, `Notification`, `SessionStart`, `SessionEnd`.
-
 ---
 
 ## Important Notice

package/bin/rune.js

@@ -656,8 +656,9 @@ function openRune(file) {
 
 function runRune(file, restArgs) {
   if (!file) {
-    console.log('Usage: rune run <file.rune> "prompt" [--output json|text]')
+    console.log('Usage: rune run <file.rune> "prompt" [--auto] [--output json|text]')
     console.log('Example: rune run reviewer.rune "Review the latest commit"')
+    console.log('         rune run coder.rune "Build a REST API server" --auto')
     process.exit(1)
   }
 
@@ -670,10 +671,13 @@ function runRune(file, restArgs) {
   // Parse args: prompt and flags
   let prompt = ''
   let outputFormat = 'text'
+  let autoMode = false
   for (let i = 0; i < restArgs.length; i++) {
     if (restArgs[i] === '--output' && restArgs[i + 1]) {
       outputFormat = restArgs[i + 1]
       i++
+    } else if (restArgs[i] === '--auto') {
+      autoMode = true
     } else if (!prompt) {
       prompt = restArgs[i]
     }
@@ -710,7 +714,76 @@ function runRune(file, restArgs) {
 
   const systemPrompt = systemParts.join('\n')
 
-  // Build claude CLI args
+  // Auto mode: agent can read/write files, run commands, fix errors autonomously
+  if (autoMode) {
+    console.log(`🔮 [auto] ${rune.name} is working on: ${prompt}\n`)
+
+    const claudeArgs = ['-p', '--print',
+      '--permission-mode', 'auto',
+      '--output-format', 'stream-json',
+      '--include-hook-events',
+    ]
+    if (systemPrompt) {
+      claudeArgs.push('--system-prompt', systemPrompt)
+    }
+    claudeArgs.push(prompt)
+
+    const child = spawn('claude', claudeArgs, {
+      cwd: folderPath,
+      stdio: ['pipe', 'pipe', 'pipe'],
+      env: { ...process.env },
+    })
+
+    let fullOutput = ''
+
+    child.stdout.on('data', (data) => {
+      const lines = data.toString().split('\n').filter(Boolean)
+      for (const line of lines) {
+        try {
+          const event = JSON.parse(line)
+
+          if (event.type === 'assistant' && event.subtype === 'tool_use') {
+            const tool = event.tool_name || event.name || 'unknown'
+            const input = event.input || {}
+            if (tool === 'Bash') {
+              console.log(`  ▶ Bash: ${(input.command || '').slice(0, 120)}`)
+            } else if (tool === 'Write') {
+              console.log(`  ▶ Write: ${input.file_path || ''}`)
+            } else if (tool === 'Edit') {
+              console.log(`  ▶ Edit: ${input.file_path || ''}`)
+            } else if (tool === 'Read') {
+              console.log(`  ▶ Read: ${input.file_path || ''}`)
+            } else {
+              console.log(`  ▶ ${tool}`)
+            }
+          }
+
+          if (event.type === 'result') {
+            fullOutput = event.result || ''
+            console.log(`\n${fullOutput}`)
+          }
+        } catch {}
+      }
+    })
+
+    child.stderr.on('data', (d) => { process.stderr.write(d) })
+
+    child.on('close', (code) => {
+      // Save to history
+      rune.history = rune.history || []
+      rune.history.push({ role: 'user', text: prompt, ts: Date.now() })
+      rune.history.push({ role: 'assistant', text: fullOutput.trim(), ts: Date.now() })
+      fs.writeFileSync(filePath, JSON.stringify(rune, null, 2))
+
+      if (code !== 0) console.error(`\n  ⚠️  Agent exited with code ${code}`)
+      else console.log(`\n✓ ${rune.name} finished`)
+      process.exit(code || 0)
+    })
+
+    return
+  }
+
+  // Normal mode: print-only, no tool execution
   const claudeArgs = ['-p', '--print']
   if (systemPrompt) {
     claudeArgs.push('--system-prompt', systemPrompt)
@@ -1133,6 +1206,7 @@ Usage:
     --role "description"    Set the agent's role
   rune open <file.rune>     Open a .rune file (desktop GUI)
   rune run <file.rune> "prompt"   Run agent headlessly (no GUI)
+    --auto                  Auto mode: agent writes files, runs commands, fixes errors
     --output json|text      Output format (default: text)
   rune pipe <a.rune> <b.rune> ... "prompt"   Chain agents in a pipeline
     --output json|text      Output format (default: text)

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "openrune",
-  "version": "0.2.1",
+  "version": "0.3.0",
   "description": "Rune — File-based AI Agent Harness for Claude Code",
   "keywords": ["ai", "agent", "claude", "desktop", "electron", "mcp", "claude-code", "harness", "automation"],
   "repository": {