pipeline-iq 0.1.2

A high-performance MCP server for CI/CD failure diagnosis

🚀 PipelineIQ: CI/CD Intelligence Layer

PipelineIQ is a high-performance Model Context Protocol (MCP) server that empowers LLMs to diagnose CI/CD failures with surgical precision. By bridging the gap between raw logs and actionable insights, PipelineIQ slashes MTTR and eliminates context-switching fatigue for modern DevOps teams. 🛠️

---

⚡ TL;DR

Stop hunting through Jenkins console output. PipelineIQ gives your LLM (Claude, GPT) direct, structured access to Jenkins logs, offering automated Failure Pattern Detection and Sensitive Data Redaction to get your builds back to green faster and safer. 🚀

---

✨ Core Features

🔍 Advanced Log Pattern Matching

PipelineIQ doesn't just "read" logs—it analyzes them across a vast tech stack:

• Liferay / OSGi: Bundle resolution and Blade CLI deployment errors.
• Java / Spring Boot: Compilation, dependency conflicts, and bean creation failures.
• Node.js / React: ESLint violations, Vite/Webpack compilation, and Hydration mismatches.
• Python / Django / FastAPI: Import errors, migration conflicts, and Pydantic validation.
• Infrastructure: Out-of-memory (OOM) kills, timeouts, and Docker build failures.

🛡️ Built-in Security (Log Redaction)

PipelineIQ automatically redacts sensitive information (passwords, API tokens, keys) from build logs before they are sent to the LLM, ensuring your secrets stay private.

🤖 Industry-Standard MCP

Built on FastMCP, PipelineIQ follows the latest Model Context Protocol standards, providing:

• Recursive Job Discovery: Navigate deep Jenkins folder structures with ease.
• Standardized Tools: list_jobs, list_builds, get_build_info, get_build_log, analyze_failure, and get_suggestions.
• Diagnostic Prompt: A pre-defined debug-failure prompt to guide AI agents.
• Patterns Resource: Access internal failure patterns via patterns://config.

🛠️ Tool Reference

Tool	Purpose
list_jobs	Discover Jenkins jobs and folders (sorted by latest activity).
list_builds	List history for a specific job.
get_build_info	Get metadata (status, duration, timestamp) for a build.
get_build_log	Fetch and sanitize console output.
analyze_failure	Detect known failure patterns in logs.
get_suggestions	Get actionable fix instructions for detected patterns.

---

🚀 Quick Start (using uvx)

No installation required! Run PipelineIQ instantly using uvx:

uvx pipeline-iq

---

🔑 Jenkins API Token Setup

Follow these steps to generate a secure API token:

1. Generate Token:

   • Log in to your Jenkins dashboard.
   • Click on your Username (top right) > Configure.
   • Locate the API Token section and click Add new Token.
   • Provide a name (e.g., PipelineIQ) and click Generate.
   • IMPORTANT: Copy the token immediately; it will not be shown again.

2. Required Permissions: PipelineIQ requires the following Read-Only permissions (using Matrix Authorization or Role-Based Strategy):

   • Overall: Read
   • Job: Read, Extended Read (to view logs)
   • View: Read

[!TIP] Always follow the Principle of Least Privilege. Use a dedicated service account if possible rather than your personal admin account.

---

⚙️ Installation & Configuration

Recommended: Install via pip

pip install pipeline-iq

From Source

# Clone the repo
git clone https://github.com/dev-joshi-ops/PipelineIQ.git

# Initialize virtual environment
python -m venv venv
source venv/bin/activate  # Windows: .\venv\Scripts\activate

# Install dependencies (Editable Mode)
pip install -e .

# OR run instantly using uv (no manual install needed)
uv run pipeline-iq

Environment Configuration 🔒

Create a .env file in the root directory or export these variables:

Variable	Required	Default	Description
JENKINS_URL	Yes	—	Base URL of Jenkins server
JENKINS_USER	Yes	—	Jenkins username
JENKINS_TOKEN	Yes	—	Jenkins API token (read-only)
LOG_LEVEL	No	INFO	Logging verbosity (DEBUG, INFO, etc.)
MAX_LOG_LINES	No	2000	Max lines to retrieve from build log
REQUEST_TIMEOUT	No	30	HTTP request timeout in seconds

Example .env:

JENKINS_URL=https://jenkins.yourdomain.com
JENKINS_USER=admin
JENKINS_TOKEN=your_api_token_here

---

🔌 Client Configuration (Recommended: uvx)

The easiest way to use PipelineIQ is with uvx.

1. Claude Desktop

Edit your claude_desktop_config.json (Windows: %APPDATA%\Claude\claude_desktop_config.json):

{
  "mcpServers": {
    "pipeline-iq": {
      "command": "uvx",
      "args": ["pipeline-iq"],
      "env": {
        "JENKINS_URL": "https://your-jenkins-url",
        "JENKINS_USER": "your-username",
        "JENKINS_TOKEN": "your-api-token"
      }
    }
  }
}

2. Antigravity

Edit your mcp_config.json (Windows: %APPDATA%\.gemini\antigravity\mcp_config.json):

{
  "mcpServers": {
    "pipeline-iq": {
      "command": "uvx",
      "args": ["pipeline-iq"],
      "env": {
        "JENKINS_URL": "https://your-jenkins-url",
        "JENKINS_USER": "your-username",
        "JENKINS_TOKEN": "your-api-token"
      }
    }
  }
}

---

⚙️ Development Setup (Manual)

If you want to contribute or run from source:

# Clone the repo
git clone https://github.com/dev-joshi-ops/PipelineIQ.git

# Initialize virtual environment
python -m venv venv
source venv/bin/activate  # Windows: .\venv\Scripts\activate

# Install dependencies (Editable Mode)
pip install -e .

---

✍️ About the Author

Built with <3 by Dev Joshi

---

📄 License

MIT © 2026 PipelineIQ