Delegates agentic sessions to Kubernetes-hosted Claude agents running on the Ambient Code Platform, enabling creation and management of AI agent sessions, sending messages to sessions, and monitoring session status through OpenShift/Kubernetes cluster integration.
Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@Ambient Code Platform MCP Serverlist my agentic sessions in the development project"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
MCP ACP Server
A Model Context Protocol (MCP) server for managing Ambient Code Platform (ACP) sessions on OpenShift/Kubernetes clusters.
Quick Start
Get started in 5 minutes:
First Command: Use acp_whoami to check my authentication
Overview
This MCP server provides 19 comprehensive tools for interacting with the Ambient Code Platform, enabling:
Session Management: List, create, delete, restart, clone sessions
Bulk Operations: Efficiently manage multiple sessions at once
Advanced Filtering: Filter by status, age, display name, and more
Debugging: Retrieve logs, transcripts, and metrics
Cluster Management: Multi-cluster support with easy switching
Safety First: Dry-run mode on all mutating operations
Security: This server implements comprehensive security measures including input validation, command injection prevention, timeout controls, and resource limits. See SECURITY.md for details.
Features
acp_add_cluster: Add new cluster configurations
acp_bulk_delete_sessions: Delete multiple sessions at once
acp_bulk_stop_sessions: Stop multiple running sessions
acp_clone_session: Clone existing session configurations
acp_create_session_from_template: Create sessions from predefined templates
acp_delete_session: Delete sessions with dry-run preview
acp_export_session: Export session data for archival
acp_get_session_logs: Retrieve container logs for debugging
acp_get_session_metrics: Get usage statistics and analytics
acp_get_session_transcript: Retrieve conversation history
acp_list_clusters: List configured cluster aliases
acp_list_sessions: Enhanced filtering by status, display name, age, and sorting
acp_list_workflows: Discover available workflows
acp_login: Web-based authentication flow
acp_restart_session: Restart stopped sessions
acp_switch_cluster: Switch between configured clusters
acp_update_session: Update session metadata
acp_whoami: Check authentication status
Dry-Run Mode: All mutating operations support a dry_run parameter for safe preview before executing.
Installation
From PyPI (when published)
From Source
Requirements:
Python 3.10+
OpenShift CLI (
oc) installed and in PATHAccess to an OpenShift cluster with ACP
See USAGE_GUIDE.md for detailed installation instructions.
Configuration
1. Create Cluster Configuration
Create ~/.config/acp/clusters.yaml:
2. Configure MCP Client
For Claude Desktop, edit your configuration file:
macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json
Linux: ~/.config/claude/claude_desktop_config.json
Add the ACP server:
3. Authenticate with OpenShift
Usage Examples
List Sessions with Filtering
Delete Session with Dry-Run
Bulk Operations
Get Session Logs
See USAGE_GUIDE.md for 40+ detailed examples and workflow patterns.
Tool Reference
For complete API specifications, see API_REFERENCE.md.
Quick Reference
Tool | Description |
| List/filter sessions with advanced options |
| Delete session with dry-run support |
| Restart stopped sessions |
| Delete multiple sessions |
| Stop multiple sessions |
| Get container logs |
| List configured clusters |
| Check authentication status |
| Clone session configuration |
| Get conversation history |
| Update session metadata |
| Export session data |
| Get usage statistics |
| Discover workflows |
| Create from template |
| Authenticate to cluster |
| Switch cluster context |
| Add cluster to config |
Architecture
The server is built using:
MCP SDK: Standard MCP protocol implementation
OpenShift CLI: Underlying
occommands for ACP operationsAsync I/O: Non-blocking operations for performance
YAML Configuration: Flexible cluster management
See ARCHITECTURE.md for complete system design.
Security
This server implements defense-in-depth security:
Input Validation: DNS-1123 format validation for all resource names
Command Injection Prevention: Secure subprocess execution (never shell=True)
Resource Exhaustion Protection: Timeouts and limits on all operations
Secure Temporary Files: Random prefixes, 0600 permissions
Path Traversal Prevention: Configuration and workflow file validation
Resource Type Whitelist: Only agenticsession, pods, event resources
Sensitive Data Filtering: Tokens/passwords removed from logs
See SECURITY.md for complete security documentation.
Development
Running Tests
Code Quality
See DEVELOPMENT.md for contributing guidelines.
Documentation
USAGE_GUIDE.md - Complete usage guide with examples
API_REFERENCE.md - Full API specifications for all 19 tools
ARCHITECTURE.md - System architecture and design
SECURITY.md - Security features and best practices
DEVELOPMENT.md - Development and contributing guide
CLEANROOM_SPEC.md - Re-implementation specification
Roadmap
Current implementation provides all planned features (19 tools). Future enhancements may include:
Rate Limiting: Per-client request limits for HTTP exposure
Audit Logging: Structured audit trail and SIEM integration
Enhanced Authentication: OAuth2/OIDC support, MFA
Network Security: mTLS for MCP transport, certificate pinning
Advanced Metrics: Cost analysis, performance tracking
See the GitHub issue tracker for planned features and community requests.
Contributing
Contributions are welcome! Please:
Fork the repository
Create a feature branch
Add tests for new functionality
Ensure all tests pass (
pytest)Ensure code quality checks pass (
make check)Submit a pull request
See DEVELOPMENT.md for detailed guidelines.
License
MIT License - See LICENSE file for details
Support
For issues and feature requests, please use the GitHub issue tracker.
For usage questions, see:
USAGE_GUIDE.md - Complete usage guide
API_REFERENCE.md - API specifications
SECURITY.md - Security features
Status
Code: ✅ Production-Ready Tests: ✅ All Passing (13/13 security tests) Documentation: ✅ Complete Security: ✅ Hardened with defense-in-depth Tools: ✅ 19 tools fully implemented
Ready for production use 🚀