juno-code 1.0.38 → 1.0.41

package/README.md

@@ -270,6 +270,99 @@ Then run with the hook:
 ./.juno_task/scripts/run_until_completion.sh --pre-run-hook SLACK_SYNC -s claude -i 5 -v
 ```
 
+## GitHub Integration
+
+juno-code includes built-in GitHub integration for issue tracking and automated responses. The system monitors GitHub repositories, creates kanban tasks from issues, and posts agent responses as threaded comments with automatic issue closure.
+
+### How It Works
+
+1. **Fetch**: `github.py fetch` monitors a GitHub repository and creates kanban tasks from new issues
+2. **Process**: The AI agent processes tasks and records responses in the kanban
+3. **Respond**: `github.py respond` posts agent responses as comments on GitHub issues and closes them
+
+This enables a workflow where team members can submit tasks via GitHub issues and receive AI-generated responses with automatic issue closure.
+
+### Setup
+
+1. **Create a GitHub Personal Access Token**:
+   - Go to https://github.com/settings/tokens and create a new token (classic)
+   - Grant these permissions:
+     - `repo` (full control of private repositories)
+     - `public_repo` (access to public repositories)
+   - Copy the token (starts with `ghp_`)
+
+2. **Configure Environment**:
+   ```bash
+   # In project root .env file
+   GITHUB_TOKEN=ghp_your_token_here
+   GITHUB_REPO=owner/repo  # Optional default repository
+   GITHUB_LABELS=bug,priority  # Optional label filter
+   ```
+
+3. **Usage**:
+   ```bash
+   # Fetch issues from GitHub and create tasks
+   ./.juno_task/scripts/github.py fetch --repo owner/repo
+
+   # Filter by labels
+   ./.juno_task/scripts/github.py fetch --repo owner/repo --labels bug,priority
+
+   # Post completed task responses back to GitHub
+   ./.juno_task/scripts/github.py respond --tag github-issue
+
+   # Bidirectional sync (fetch + respond)
+   ./.juno_task/scripts/github.py sync --repo owner/repo
+
+   # Continuous sync mode with interval
+   ./.juno_task/scripts/github.py sync --repo owner/repo --continuous --interval 600
+
+   # One-time sync
+   ./.juno_task/scripts/github.py sync --repo owner/repo --once
+
+   # Dry run to preview what would be posted
+   ./.juno_task/scripts/github.py respond --dry-run --verbose
+   ```
+
+### Key Features
+
+- **Tag-based identification**: Uses `github_issue_owner_repo_123` format for O(1) lookups
+- **State tracking**: Maintains state in `.juno_task/github/state.ndjson` and `responses.ndjson`
+- **Automatic closure**: Issues are automatically closed after posting the agent response
+- **Commit linking**: Includes commit hash in comment if available
+- **Response format**: Posts `agent_response` field from completed kanban tasks
+
+### Automated GitHub Workflow with Hooks
+
+Use the `--pre-run` flag to sync with GitHub before each juno-code run:
+
+```bash
+# Fetch GitHub issues before starting work
+./.juno_task/scripts/run_until_completion.sh \
+  --pre-run "./.juno_task/scripts/github.py fetch --repo owner/repo" \
+  -s claude -i 5 -v
+```
+
+Or configure hooks in `.juno_task/config.json`:
+
+```json
+{
+  "hooks": {
+    "GITHUB_SYNC": {
+      "commands": [
+        "./.juno_task/scripts/github.py fetch --repo owner/repo --labels bug",
+        "./.juno_task/scripts/github.py respond --tag github-issue"
+      ]
+    }
+  }
+}
+```
+
+Then run with the hook:
+
+```bash
+./.juno_task/scripts/run_until_completion.sh --pre-run-hook GITHUB_SYNC -s claude -i 5 -v
+```
+
 ## run_until_completion.sh
 
 The `run_until_completion.sh` script continuously runs juno-code until all kanban tasks are completed. It uses a do-while loop pattern: juno-code runs at least once, then continues while tasks remain in backlog, todo, or in_progress status.

package/dist/bin/cli.js

@@ -1337,7 +1337,7 @@ var init_config = __esm({
     DEFAULT_CONFIG = {
       // Core settings
       defaultSubagent: "claude",
-      defaultBackend: "mcp",
+      defaultBackend: "shell",
       defaultMaxIterations: 50,
       // Logging settings
       logLevel: "info",
@@ -13426,8 +13426,11 @@ var init_script_installer = __esm({
         // Wrapper script for Slack fetch
         "slack_respond.py",
         // Core logic for sending responses to Slack
-        "slack_respond.sh"
+        "slack_respond.sh",
         // Wrapper script for Slack respond
+        // GitHub integration script (single-file architecture)
+        "github.py"
+        // Unified GitHub integration (fetch, respond, sync)
       ];
       /**
        * Get the templates scripts directory from the package
@@ -14044,15 +14047,17 @@ async function validateStartupConfigs(baseDir = process.cwd(), verbose = false)
     console.error(chalk15__default.default.blue("\u{1F50D} Validating JSON configuration files...\n"));
   }
   try {
-    let backendType = "mcp";
+    let backendType = "shell";
+    let backendSetViaCliArg = false;
     const backendArgIndex = process.argv.findIndex((arg) => arg === "-b" || arg === "--backend");
     if (backendArgIndex !== -1 && process.argv[backendArgIndex + 1]) {
       const cliBackend = process.argv[backendArgIndex + 1].toLowerCase().trim();
       if (cliBackend === "shell" || cliBackend === "mcp") {
         backendType = cliBackend;
+        backendSetViaCliArg = true;
       }
     }
-    if (backendType === "mcp") {
+    if (!backendSetViaCliArg) {
       const envBackend = process.env.JUNO_CODE_AGENT || process.env.JUNO_CODE_BACKEND || process.env.JUNO_TASK_BACKEND;
       if (envBackend) {
         const normalized = envBackend.toLowerCase().trim();
@@ -14551,22 +14556,19 @@ and for each part create a seperate .md file under @.juno_task/spec/*
 
 ## ULTIMATE Goal
 We want to achieve the main Task with respect to the Constraints section
+
+Part 1)
 Consider missing steps and plan. If the step is missing then author the specification at @.juno_task/spec/FILENAME.md (do NOT assume that it does not exist, search before creating). The naming of the module should be GenZ named and not conflict with another module name. If you create a new step then document the plan to implement in @.juno_task/plan.md
 
 
+Part 2) after completing the plan, and spec, create task for implementing each part on kanban './.juno_task/scripts/kanban.sh' You need to create a task for each step of implementation and testing. You need to go through the project, the spec and plan at the end to make sure you have covered all tasks on the kanban. We will later on implement tasks from kanban one by one.
+After completing the proccess an implementer agent would start the job and go through kanban tasks one by one.
+
+
 ### Constraints
 **Preferred Subagent**: {{SUBAGENT}}
 **Repository URL**: {{GIT_URL}}
-
-
-## Environment Setup
-[Empty]
-
-### 2. Package Installation
-[Empty]
-
-### 3. Test Installation
-[Empty]`,
+`,
         variables: [
           {
             name: "main_task",
@@ -14911,6 +14913,23 @@ Items that have been resolved will be moved here.`,
 **Git Repository:** {{GIT_URL}}
 **Configuration Date:** {{CURRENT_DATE}}
 
+## Kanban Task Management
+
+\`\`\`bash
+# List tasks
+./.juno_task/scripts/kanban.sh list --limit 5 --sort asc 
+./.juno_task/scripts/kanban.sh list --status [backlog|todo|in_progress|done] --sort asc
+
+# Task operations
+./.juno_task/scripts/kanban.sh get {TASK_ID}
+./.juno_task/scripts/kanban.sh mark [in_progress|done|todo] --id {TASK_ID} --response "message"
+./.juno_task/scripts/kanban.sh update {TASK_ID} --commit {COMMIT_HASH}
+\`\`\`
+
+When a task on kanban, has related_tasks key, you need to get the task to understand the complete picture of tasks related to the current current task, you can get all the context through
+\`./.juno_task/scripts/kanban.sh get {TASK_ID}\`
+
+
 ## Agent-Specific Instructions
 
 ### {{SUBAGENT}} Configuration
@@ -15025,6 +15044,24 @@ python -m pytest tests/ --cov=src --cov-report=term-missing
 - **Code Quality:** Focus on production-ready, well-documented code
 - **Testing:** Comprehensive unit and integration tests required
 
+## Kanban Task Management
+
+\`\`\`bash
+# List tasks
+./.juno_task/scripts/kanban.sh list --limit 5 --sort asc 
+./.juno_task/scripts/kanban.sh list --status [backlog|todo|in_progress|done] --sort asc
+
+# Task operations
+./.juno_task/scripts/kanban.sh get {TASK_ID}
+./.juno_task/scripts/kanban.sh mark [in_progress|done|todo] --id {TASK_ID} --response "message"
+./.juno_task/scripts/kanban.sh update {TASK_ID} --commit {COMMIT_HASH}
+\`\`\`
+
+When a task on kanban, has related_tasks key, you need to get the task to understand the complete picture of tasks related to the current current task, you can get all the context through
+\`./.juno_task/scripts/kanban.sh get {TASK_ID}\`
+
+
+
 ## Build & Test Commands
 
 **Environment Setup:**
@@ -24307,7 +24344,7 @@ function handleCLIError(error, verbose = false) {
   process.exit(EXIT_CODES.UNEXPECTED_ERROR);
 }
 function setupGlobalOptions(program) {
-  program.option("-v, --verbose", "Enable verbose output with detailed progress").option("-q, --quiet", "Disable rich formatting, use plain text").option("-c, --config <path>", "Configuration file path (.json, .toml, pyproject.toml)").option("-l, --log-file <path>", "Log file path (auto-generated if not specified)").option("--no-color", "Disable colored output").option("--log-level <level>", "Log level for output (error, warn, info, debug, trace)", "info").option("-s, --subagent <name>", "Subagent to use (claude, cursor, codex, gemini)").option("-b, --backend <type>", "Backend to use (mcp, shell)").option("-m, --model <name>", "Model to use (subagent-specific)").option("--agents <config>", "Agents configuration (forwarded to shell backend, ignored for MCP)").option("--tools <tools...>", 'Specify the list of available tools from the built-in set (only works with --print mode). Use "" to disable all tools, "default" to use all tools, or specify tool names (e.g. "Bash,Edit,Read"). Passed to shell backend, ignored for MCP.').option("--allowed-tools <tools...>", 'Permission-based filtering of specific tool instances (e.g. "Bash(git:*) Edit"). Default when not specified: Task, Bash, Glob, Grep, ExitPlanMode, Read, Edit, Write, NotebookEdit, WebFetch, TodoWrite, WebSearch, BashOutput, KillShell, Skill, SlashCommand, EnterPlanMode. Passed to shell backend, ignored for MCP.').option("--disallowed-tools <tools...>", "Disallowed tools for Claude (passed to shell backend, ignored for MCP). By default, no tools are disallowed").option("--append-allowed-tools <tools...>", "Append tools to the default allowed-tools list (mutually exclusive with --allowed-tools). Passed to shell backend, ignored for MCP.").option("--mcp-timeout <number>", "MCP server timeout in milliseconds", parseInt).option("--enable-feedback", "Enable interactive feedback mode (F+Enter to enter, Q+Enter to submit)").option("-r, --resume <sessionId>", "Resume a conversation by session ID (shell backend only)").option("--continue", "Continue the most recent conversation (shell backend only)");
+  program.option("-v, --verbose", "Enable verbose output with detailed progress").option("-q, --quiet", "Disable rich formatting, use plain text").option("-c, --config <path>", "Configuration file path (.json, .toml, pyproject.toml)").option("-l, --log-file <path>", "Log file path (auto-generated if not specified)").option("--no-color", "Disable colored output").option("--log-level <level>", "Log level for output (error, warn, info, debug, trace)", "info").option("-s, --subagent <name>", "Subagent to use (claude, cursor, codex, gemini)").option("-b, --backend <type>", "Backend to use (mcp, shell) - default: shell").option("-m, --model <name>", "Model to use (subagent-specific)").option("--agents <config>", "Agents configuration (forwarded to shell backend, ignored for MCP)").option("--tools <tools...>", 'Specify the list of available tools from the built-in set (only works with --print mode). Use "" to disable all tools, "default" to use all tools, or specify tool names (e.g. "Bash,Edit,Read"). Passed to shell backend, ignored for MCP.').option("--allowed-tools <tools...>", 'Permission-based filtering of specific tool instances (e.g. "Bash(git:*) Edit"). Default when not specified: Task, Bash, Glob, Grep, ExitPlanMode, Read, Edit, Write, NotebookEdit, WebFetch, TodoWrite, WebSearch, BashOutput, KillShell, Skill, SlashCommand, EnterPlanMode. Passed to shell backend, ignored for MCP.').option("--disallowed-tools <tools...>", "Disallowed tools for Claude (passed to shell backend, ignored for MCP). By default, no tools are disallowed").option("--append-allowed-tools <tools...>", "Append tools to the default allowed-tools list (mutually exclusive with --allowed-tools). Passed to shell backend, ignored for MCP.").option("--mcp-timeout <number>", "MCP server timeout in milliseconds", parseInt).option("--enable-feedback", "Enable interactive feedback mode (F+Enter to enter, Q+Enter to submit)").option("-r, --resume <sessionId>", "Resume a conversation by session ID (shell backend only)").option("--continue", "Continue the most recent conversation (shell backend only)").option("--til-completion", "Run juno-code in a loop until all kanban tasks are complete (aliases: --until-completion, --run-until-completion, --till-complete)").option("--until-completion", "Alias for --til-completion").addOption(new commander.Option("--run-until-completion", "Alias for --til-completion").hideHelp()).addOption(new commander.Option("--till-complete", "Alias for --til-completion").hideHelp()).option("--pre-run-hook <hooks...>", "Execute named hooks from .juno_task/config.json before each iteration (only with --til-completion)");
   program.exitOverride((err) => {
     if (err.code === "commander.helpDisplayed") {
       process.exit(0);
@@ -24332,6 +24369,61 @@ function setupMainCommand(program) {
         Object.entries(globalOptions).filter(([_, v]) => v !== void 0)
       );
       const allOptions2 = { ...definedGlobalOptions, ...options };
+      if (allOptions2.tilCompletion || allOptions2.untilCompletion || allOptions2.runUntilCompletion || allOptions2.tillComplete) {
+        const { spawn: spawn4 } = await import('child_process');
+        const path24 = await import('path');
+        const fs22 = await import('fs-extra');
+        const scriptPath = path24.join(process.cwd(), ".juno_task", "scripts", "run_until_completion.sh");
+        if (!await fs22.pathExists(scriptPath)) {
+          console.error(chalk15__default.default.red.bold("\n\u274C Error: run_until_completion.sh not found"));
+          console.error(chalk15__default.default.red(`   Expected location: ${scriptPath}`));
+          console.error(chalk15__default.default.yellow('\n\u{1F4A1} Suggestion: Run "juno-code init" to initialize the project'));
+          process.exit(1);
+        }
+        const scriptArgs = [];
+        if (allOptions2.preRunHook && Array.isArray(allOptions2.preRunHook)) {
+          for (const hook of allOptions2.preRunHook) {
+            scriptArgs.push("--pre-run-hook", hook);
+          }
+        }
+        const completionFlags = ["--til-completion", "--until-completion", "--run-until-completion", "--till-complete"];
+        const forwardedArgs = process.argv.slice(2).filter(
+          (arg) => !completionFlags.includes(arg) && !arg.startsWith("--pre-run-hook")
+        );
+        scriptArgs.push(...forwardedArgs);
+        const child = spawn4(scriptPath, scriptArgs, {
+          stdio: "inherit",
+          cwd: process.cwd()
+        });
+        process.removeAllListeners("SIGINT");
+        process.removeAllListeners("SIGTERM");
+        let childExited = false;
+        const signalHandler = (signal) => {
+          if (!childExited && child.pid) {
+            try {
+              process.kill(child.pid, signal);
+            } catch (err) {
+            }
+          }
+        };
+        process.on("SIGINT", signalHandler);
+        process.on("SIGTERM", signalHandler);
+        child.on("exit", (code) => {
+          childExited = true;
+          process.removeListener("SIGINT", signalHandler);
+          process.removeListener("SIGTERM", signalHandler);
+          process.exit(code || 0);
+        });
+        child.on("error", (error) => {
+          childExited = true;
+          process.removeListener("SIGINT", signalHandler);
+          process.removeListener("SIGTERM", signalHandler);
+          console.error(chalk15__default.default.red.bold("\n\u274C Error executing run_until_completion.sh"));
+          console.error(chalk15__default.default.red(`   ${error.message}`));
+          process.exit(1);
+        });
+        return;
+      }
       if (!globalOptions.subagent && !options.prompt && !options.interactive && !options.interactivePrompt) {
         const fs22 = await import('fs-extra');
         const path24 = await import('path');
@@ -24572,9 +24664,12 @@ ${chalk15__default.default.blue.bold("\u{1F3AF} Juno Code")} - TypeScript CLI fo
 `);
   program.addHelpText("afterAll", `
 ${chalk15__default.default.blue.bold("Examples:")}
-  ${chalk15__default.default.gray("# Initialize new project")}
+  ${chalk15__default.default.gray("# Initialize new project (interactive mode)")}
   juno-code init
 
+  ${chalk15__default.default.gray("# Initialize with inline mode (automation-friendly)")}
+  juno-code init "Build a REST API" --subagent claude --git-repo https://github.com/user/repo
+
   ${chalk15__default.default.gray("# Start execution using .juno_task/init.md")}
   juno-code start
 
@@ -24604,7 +24699,7 @@ ${chalk15__default.default.blue.bold("Examples:")}
   juno-code config create development
 
   ${chalk15__default.default.gray("# Setup Git repository")}
-  juno-code setup-git https://github.com/owner/repo
+  juno-code setup-git https://github.com/askbudi/juno-code
 
 ${chalk15__default.default.blue.bold("Environment Variables:")}
   JUNO_CODE_SUBAGENT              Default subagent (claude, cursor, codex, gemini)
@@ -24627,8 +24722,9 @@ ${chalk15__default.default.blue.bold("Configuration:")}
   4. Built-in defaults (lowest priority)
 
 ${chalk15__default.default.blue.bold("Support:")}
-  Documentation: https://github.com/owner/juno-code#readme
-  Issues: https://github.com/owner/juno-code/issues
+  Documentation: https://github.com/askbudi/juno-code#readme
+  Issues: https://github.com/askbudi/juno-code/issues
+  Website: https://askbudi.ai
   License: MIT
 
 `);