claude-intern 1.0.1 → 1.1.2

package/.env.sample

@@ -41,8 +41,5 @@ CLAUDE_CLI_PATH=claude
 
 # Note: Bitbucket workspace is automatically detected from your git remote URL
 
-# Optional: JIRA Status Transition After PR Creation
-# Set the status to transition the JIRA task to after successfully creating a PR
-# Common values: "In Review", "Code Review", "Under Review", "Ready for Review"
-# Leave empty or comment out to skip status transition
-# JIRA_PR_STATUS=In Review
+# Note: JIRA PR status transitions are now configured per-project in .claude-intern/settings.json
+# See settings.json for configuring different PR statuses for different JIRA projects

package/README.md

@@ -42,27 +42,49 @@ bun run install-global
 
 ## Configuration
 
+### Quick Setup with Init Command
+
+The easiest way to set up claude-intern for your project is using the `init` command:
+
+```bash
+# Initialize project-specific configuration
+claude-intern init
+```
+
+This creates a `.claude-intern` folder in your current project with:
+- `.env` - Your project-specific configuration file with JIRA credentials
+- `.env.sample` - Template with all configuration options
+- `settings.json` - Per-project settings (PR status transitions, etc.)
+
+**Automatic .gitignore Protection:** The `init` command automatically adds `.claude-intern/.env` to your `.gitignore` file (or creates one if it doesn't exist) to prevent accidentally committing credentials to version control.
+
+After running `init`:
+1. Edit `.claude-intern/.env` with your JIRA credentials
+2. (Optional) Edit `.claude-intern/settings.json` to configure per-project PR status transitions
+
 ### Environment File Setup
 
 The tool searches for `.env` files in the following order:
 
 1. **Custom path** (if specified with `--env-file`)
-2. **Current working directory** (where you run the command)
-3. **User home directory** (`~/.env`)
-4. **Tool installation directory**
+2. **Project-specific** (`.claude-intern/.env` - recommended)
+3. **Current working directory** (`.env`)
+4. **User home directory** (`~/.env`)
+5. **Tool installation directory**
 
-For global installation, it's recommended to either:
-- Place `.env` in your project directory (most common)
-- Place `.env` in your home directory (`~/.env`) for global access
+Configuration options:
 
 ```bash
 # Option 1: Project-specific (recommended)
+claude-intern init  # Creates .claude-intern/.env
+
+# Option 2: Current directory
 cp .env.sample .env
 
-# Option 2: Global configuration
+# Option 3: Global configuration
 cp .env.sample ~/.env
 
-# Option 3: Custom location
+# Option 4: Custom location
 claude-intern PROJ-123 --env-file /path/to/custom.env
 ```
 
@@ -77,10 +99,35 @@ Update your `.env` file with your JIRA details:
    Optional PR integration:
    - `GITHUB_TOKEN`: GitHub personal access token for creating pull requests
    - `BITBUCKET_TOKEN`: Bitbucket app password for creating pull requests
-   - `JIRA_PR_STATUS`: Auto-transition JIRA status after PR creation (e.g., "In Review")
 
    The `.env.sample` file includes helpful comments and optional configuration options.
 
+   **Note:** JIRA PR status transitions are now configured per-project in `settings.json` (see below).
+
+### Per-Project Settings (settings.json)
+
+The `.claude-intern/settings.json` file allows you to configure project-specific behavior:
+
+```json
+{
+  "projects": {
+    "PROJ": {
+      "prStatus": "In Review"
+    },
+    "ABC": {
+      "prStatus": "Code Review"
+    }
+  }
+}
+```
+
+**Configuration options:**
+- `prStatus`: JIRA status to transition to after PR creation for a specific project
+  - Each project key can have its own status workflow
+  - If not configured, no status transition will occur
+
+**Example:** If you work with multiple JIRA projects that have different workflows (e.g., "PROJ" uses "In Review" but "ABC" uses "Code Review"), configure each project's status in `settings.json`.
+
 ## Usage
 
 ### Single Task Processing
@@ -121,6 +168,13 @@ claude-intern TASK-123 --create-pr
 
 # Create pull request targeting specific branch
 claude-intern TASK-123 --create-pr --pr-target-branch develop
+
+# Testing options - skip ALL JIRA comments (feasibility + implementation)
+claude-intern TASK-123 --skip-jira-comments
+
+# Local testing without PR (commits stay local, no push to remote)
+# Simply omit --create-pr to skip pushing to remote
+claude-intern TASK-123 --skip-jira-comments
 ```
 
 #### Local Development Usage
@@ -186,6 +240,31 @@ claude-intern --jql "\"Epic Link\" = PROJ-100" --create-pr
 claude-intern --jql "sprint in openSprints() AND assignee = currentUser()" --max-turns 500
 ```
 
+### Automated Processing with Cron
+
+You can set up automated task processing using cron jobs. This is useful for continuously picking up new tasks labeled for the intern to work on:
+
+```bash
+# Example: Process tasks labeled "Intern" in open sprints every 10 minutes
+# Add to crontab (run: crontab -e)
+*/10 * * * * cd /path/to/your/project && claude-intern --jql 'statusCategory = "To Do" AND sprint in openSprints() AND labels IN (Intern) ORDER BY created DESC' --max-turns 500 --create-pr --pr-target-branch master >> /tmp/claude-intern-cron.log 2>&1
+
+# Example: Process assigned tasks every hour
+0 * * * * cd /path/to/your/project && claude-intern --jql 'assignee = currentUser() AND status = "To Do" AND labels IN (AutoImpl)' --create-pr >> /tmp/claude-intern-cron.log 2>&1
+
+# Example: Process high-priority bugs twice daily
+0 9,17 * * * cd /path/to/your/project && claude-intern --jql 'type = Bug AND priority = High AND status = "To Do" AND labels IN (Intern)' --max-turns 300 --create-pr >> /tmp/claude-intern-cron.log 2>&1
+```
+
+**Important notes for cron setup:**
+- Always change to your project directory (`cd /path/to/your/project`) to ensure the correct `.claude-intern/.env` is loaded
+- Use absolute paths or ensure PATH includes `claude-intern` and `claude` binaries
+- Redirect output to a log file for monitoring (`>> /tmp/claude-intern-cron.log 2>&1`)
+- Use the `ORDER BY created DESC` clause to process newest tasks first
+- Consider using labels (e.g., `labels = "Intern"`) to mark tasks for automated processing
+- Test your JQL query manually before adding to cron to ensure it returns the expected tasks
+- Monitor the log file regularly to ensure the cron job is running successfully
+
 ## What it does
 
 1. Fetches the JIRA task details including:
@@ -198,13 +277,27 @@ claude-intern --jql "sprint in openSprints() AND assignee = currentUser()" --max
 
 3. Creates a feature branch named `feature/task-id` (e.g., `feature/proj-123`)
 
-4. Runs `claude -p` with enhanced permissions and extended conversation limits for automatic implementation
+4. Runs optional feasibility assessment to validate task clarity (skippable with `--skip-clarity-check`)
+   - Posts assessment results to JIRA (skippable with `--skip-jira-comments`)
+   - Saves assessment results for debugging:
+     - `{output-dir}/{task-key}/feasibility-assessment.md` (formatted results with JSON)
+     - `{output-dir}/{task-key}/feasibility-assessment-failed.txt` (raw output on parse failure)
+
+5. Runs `claude -p` with enhanced permissions and extended conversation limits for automatic implementation
 
-5. Automatically commits all changes with a descriptive commit message after Claude completes successfully
+6. Saves Claude's implementation summary to local files for analysis:
+   - `{output-dir}/{task-key}/implementation-summary.md` (successful)
+   - `{output-dir}/{task-key}/implementation-summary-incomplete.md` (incomplete/failed)
 
-6. Pushes the feature branch to remote repository (when creating PRs)
+7. Automatically commits all changes with a descriptive commit message after Claude completes successfully
 
-7. Optionally creates pull requests on GitHub or Bitbucket with detailed implementation summaries
+8. Pushes the feature branch to remote repository (when creating PRs)
+
+9. Optionally creates pull requests on GitHub or Bitbucket with detailed implementation summaries
+
+10. Posts implementation results back to JIRA as comments (skippable with `--skip-jira-comments`)
+    - Includes feasibility assessment results
+    - Includes implementation summary
 
 ## Requirements
 
@@ -216,14 +309,12 @@ claude-intern --jql "sprint in openSprints() AND assignee = currentUser()" --max
 ## Quick Start
 
 1. Install globally: `npm install -g claude-intern`
-2. Get the sample environment file:
-   ```bash
-   # Download .env.sample from the repository
-   curl -o .env https://raw.githubusercontent.com/danii1/claude-intern/master/.env.sample
-   ```
-3. Configure your JIRA credentials in `.env` (or `~/.env` for global access)
+2. Initialize your project: `claude-intern init`
+3. Configure your JIRA credentials in `.claude-intern/.env`
 4. Run from any directory: `claude-intern PROJ-123`
 
+**Alternative:** For global configuration across all projects, place `.env` in your home directory (`~/.env`)
+
 See [USAGE.md](./USAGE.md) for detailed usage scenarios and troubleshooting.  
 See [GLOBAL_USAGE.md](./GLOBAL_USAGE.md) for quick reference on global installation.  
 See [ENV_SETUP.md](./ENV_SETUP.md) for comprehensive environment configuration guide.