repo-rover-runner 0.1.17

Cross-platform GitHub/Bitbucket repository automation toolkit

Repo Rover Runner

+----------------------------+
|      REPO ROVER RUNNER     |
+----------------------------+

Repo Rover Runner is a local automation toolkit that standardizes a small set of git workflows for both GitHub and Bitbucket repositories.

It is built around a Python CLI plus Windows (.bat) and Linux/macOS (.sh) helper scripts.

Script help documents are available under docs/utilities/.

What This Workspace Is Doing

This project automates these tasks against a remote git repository:

1. Connectivity check (ping remote)
2. Clone if needed
3. Create or attach to a working branch
4. List local and remote branches
5. Copy one or more files/folders into the repo
6. Commit and push to remote branch

It does not call GitHub or Bitbucket REST APIs directly. It relies on standard git commands, so it works with either provider as long as the remote URL and credentials are valid.

Core CLI

Primary entrypoint:

python repo_rover_runner_client.py --help

Installed command aliases (Windows installs .exe launchers):

repo-rover-runner --help
rrr --help

Legacy compatibility wrapper:

python repo_rover_runner_cli.py --help

MCP Server

The project now ships an MCP stdio server that exposes the same core operations.

Start it with:

repo-rover-runner-mcp

Exposed MCP tools:

1. ping_repo
2. clone_repo
3. use_branch
4. list_branches
5. push_files

MCP Client JSON Configurations

The MCP server uses stdio transport, so most clients use a JSON entry under mcpServers.

Minimal server command:

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "repo-rover-runner-mcp"
      }
   }
}

If the command is not on PATH, use an explicit Python executable and script path.

Windows example (project-local virtual environment):

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "C:\\MyProjects\\repo-rover-runner\\.venv\\Scripts\\python.exe",
         "args": [
            "C:\\MyProjects\\repo-rover-runner\\repo_rover_runner_mcp_server.py"
         ]
      }
   }
}

Linux/macOS example:

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "/path/to/repo-rover-runner/.venv/bin/python",
         "args": [
            "/path/to/repo-rover-runner/repo_rover_runner_mcp_server.py"
         ]
      }
   }
}

Client-Specific Examples

Claude Desktop (claude_desktop_config.json):

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "repo-rover-runner-mcp",
         "env": {
            "REPO_PROVIDER": "github",
            "GITHUB_TOKEN": "${GITHUB_TOKEN}",
            "GIT_USERNAME": "x-access-token"
         }
      }
   }
}

VS Code MCP config (settings.json):

{
   "mcp": {
      "servers": {
         "repo-rover-runner": {
            "command": "C:\\MyProjects\\repo-rover-runner\\.venv\\Scripts\\python.exe",
            "args": [
               "C:\\MyProjects\\repo-rover-runner\\repo_rover_runner_mcp_server.py"
            ],
            "env": {
               "REPO_PROVIDER": "auto"
            }
         }
      }
   }
}

If repo-rover-runner-mcp is on PATH, you can also use:

{
   "mcp": {
      "servers": {
         "repo-rover-runner": {
            "command": "repo-rover-runner-mcp",
            "env": {
               "REPO_PROVIDER": "auto"
            }
         }
      }
   }
}

Cursor/Cline-style MCP config:

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "repo-rover-runner-mcp",
         "args": [],
         "env": {
            "REPO_PROVIDER": "auto"
         }
      }
   }
}

Generic host config with explicit working defaults:

{
   "mcpServers": {
      "repo-rover-runner": {
         "command": "repo-rover-runner-mcp",
         "env": {
            "REPO_PROVIDER": "bitbucket",
            "BITBUCKET_USERNAME": "your-user",
            "BITBUCKET_APP_PASSWORD": "your-app-password"
         }
      }
   }
}

Recommended Environment Variables for MCP Clients

Most clients allow per-server env values. Use those to avoid embedding credentials in URLs.

GitHub HTTPS auth:

1. REPO_PROVIDER=github
2. GITHUB_TOKEN=<token>
3. GIT_USERNAME=x-access-token (optional; default is already x-access-token)

Bitbucket HTTPS auth:

1. REPO_PROVIDER=bitbucket
2. BITBUCKET_USERNAME=<username> (or GIT_USERNAME)
3. BITBUCKET_APP_PASSWORD=<app-password> (or BITBUCKET_TOKEN)

Notes:

1. Keep secrets out of committed files.
2. Prefer environment-variable expansion supported by your MCP host.
3. If your host cannot expand ${VAR}, set values directly in that host's secure secret settings.

Supported subcommands:

1. ping
2. clone
3. use-branch
4. list-branches
5. push-files

Command Behavior Summary

ping

• Runs git ls-remote <repo-url> HEAD

clone

• Clones into destination folder
• Fails if destination exists and is not empty

use-branch

• Fetches remote first
• Reuses local branch if it exists
• Otherwise tracks remote branch if it exists
• Otherwise creates from remote base branch
• Otherwise creates from local base branch
• If repository has no commits, creates orphan branch

list-branches

• Fetches remote first
• Prints local and/or remote branches by scope: all | local | remote

push-files

• Ensures branch exists (same logic as use-branch)
• Copies provided files/folders into target directory in repo
• Stages copied paths only
• If no staged diff, exits without commit/push
• Otherwise commits and pushes with upstream tracking

Authentication Model

For SSH remotes, normal git SSH auth applies.

For HTTPS remotes, the CLI and scripts support temporary GIT_ASKPASS credentials via environment variables:

• GitHub: GITHUB_TOKEN and optional GIT_USERNAME (default x-access-token)
• Bitbucket: BITBUCKET_APP_PASSWORD (or BITBUCKET_TOKEN) and BITBUCKET_USERNAME (or GIT_USERNAME)

Provider selection:

• REPO_PROVIDER=github
• REPO_PROVIDER=bitbucket
• REPO_PROVIDER=auto (default): prefers GitHub token if present, otherwise Bitbucket credentials

The project does not persist credentials.

Configuration Files

Preferred provider-specific files:

• repo_rover_runner.github.env
• repo_rover_runner.bitbucket.env

Legacy combined template:

• repo_rover_runner.env.example
• optional runtime file: repo_rover_runner.env

Common required variables:

• REPO_URL
• LOCAL_REPO_PATH
• TARGET_BRANCH

Common optional variables:

• BASE_BRANCH (default main)
• REMOTE (default origin)
• TARGET_DIR (default integration-tests)
• COMMIT_MESSAGE
• BRANCH_TO_LIST

Script config resolution order:

1. REPO_CONFIG_FILE if provided
2. provider-specific file if REPO_PROVIDER set
3. auto-fallback: repo_rover_runner.bitbucket.env, then repo_rover_runner.github.env, then repo_rover_runner.env

Helper Scripts

The workspace includes paired scripts for Windows and Linux/macOS to run end-to-end checks quickly.

Main scripted operations:

1. test_ping_repo - remote reachability/auth check
2. create_new_branch - clone-if-needed then use-branch
3. test_list_branches - branch listing
4. test_use_branch - branch checkout/create flow
5. list_files_in_branch - file listing from a branch ref
6. test_push_dummy - push dummy_payload.txt

One-command suites:

• run_github_tests.bat / run_github_tests.sh
• run_bitbucket_tests.bat / run_bitbucket_tests.sh

Connectivity-only suites:

• run_github_connectivity.bat / run_github_connectivity.sh
• run_bitbucket_connectivity.bat / run_bitbucket_connectivity.sh

Quick Start

1. Create and fill one config file (repo_rover_runner.github.env or repo_rover_runner.bitbucket.env).
2. Set provider if needed: REPO_PROVIDER=github or REPO_PROVIDER=bitbucket.
3. Run one command:

Windows:

run_github_tests.bat

or

run_bitbucket_tests.bat

Linux/macOS:

./run_github_tests.sh

or

./run_bitbucket_tests.sh

Project Structure

Python package:

• repo_rover_runner/interfaces.py: operation contract
• repo_rover_runner/operations.py: shared git implementation and provider classes
• repo_rover_runner/factory.py: provider selection
• repo_rover_runner/auth.py: temporary HTTPS askpass setup and env resolution
• repo_rover_runner/exceptions.py: custom error type

CLI:

• repo_rover_runner_client.py: main CLI
• repo_rover_runner_cli.py: legacy wrapper to main CLI

Tests:

• tests/test_auth.py
• tests/test_factory.py
• tests/test_operations.py
• tests/test_cli_client.py

Unit Tests

Windows:

run_repo_rover_runner_unit_tests.bat

Linux/macOS:

./run_repo_rover_runner_unit_tests.sh

The test runner executes unittest with coverage and prints a coverage report.

Security Reports

Dev/security dependencies are available through either:

pip install -r requirements-dev.txt

or:

pip install ".[dev]"

Generate security reports:

Windows:

run_security_reports.bat

Linux/macOS:

./run_security_reports.sh

All generated security artifacts are written under:

1. test_results/Secutiry_reports/pip_audit_report.json
2. test_results/Secutiry_reports/bandit_report.json
3. test_results/Secutiry_reports/gitleaks_report.json
4. test_results/Secutiry_reports/pip_audit_report.html
5. test_results/Secutiry_reports/bandit_report.html
6. test_results/Secutiry_reports/gitleaks_report.html
7. test_results/Secutiry_reports/security_consolidated.html

sec-report-kit is used when available to convert JSON reports to HTML, with an automatic built-in HTML fallback to guarantee report generation.

GitHub Actions Workflows

The repository now includes these workflows under .github/workflows/:

1. ci.yml
2. publish-testpypi.yml
3. publish-pypi.yml

CI and Security

ci.yml runs on push and pull request and includes:

1. Unit tests with coverage
2. Dependency vulnerability scan with pip-audit
3. Static security scan with bandit
4. Secret scan with gitleaks

TestPyPI Publish

publish-testpypi.yml publishes to TestPyPI:

1. Manually via workflow_dispatch
2. Automatically for pre-release tags like:
   1. vX.Y.Z-rcN
   2. vX.Y.Z-betaN
   3. vX.Y.Z-alphaN

PyPI Publish

publish-pypi.yml publishes to PyPI:

1. Manually via workflow_dispatch
2. Automatically when a GitHub Release is published

Required Repository Setup for Publishing

Both publish workflows use trusted publishing (id-token: write) with pypa/gh-action-pypi-publish.

Configure these environments in GitHub repository settings:

1. testpypi
2. pypi

Then configure trusted publisher entries in TestPyPI/PyPI for this repository and workflow files.