AI Navigate

I built a CI/CD tool that auto-heals broken pipelines, runs 6 security scans, and works from your IDE via MCP

Dev.to / 3/13/2026

📰 NewsDeveloper Stack & InfrastructureTools & Practical Usage

Key Points

  • FlowEasy automates the detection and auto-healing of broken CI/CD pipelines by reading full error logs, diagnosing root causes using AI, and automatically committing fixes or opening pull requests in under 30 seconds.
  • It automatically handles common issues such as dependency version conflicts, missing environment variables, YAML syntax errors, test runner misconfigurations, and build cache invalidation.
  • FlowEasy generates secure pipelines with GitHub Actions and actively fixes problems instead of merely notifying you.
  • Its MCP Server lets you manage pipelines from your IDE (Claude Code, Cursor, Windsurf) with nine tools, including check_pipeline_status, view_logs, rerun_pipeline, rollback_deployment, analyze_failure, list_pipelines, get_pipeline_config, update_pipeline, and cancel_run.
  • A one-line MCP config sets up FlowEasy in your environment, enabling instant workflow from your editor.

Your CI/CD pipeline breaks at 2 AM. The deployment is stuck. You're half-asleep, staring at cryptic error logs, trying to figure out if it's a dependency issue, a YAML typo, or something else entirely.

I've been there too many times. That's why I built FlowEasy — a CI/CD tool that generates secure pipelines with GitHub Actions and doesn't just notify you when things break. It fixes them.

Here's what FlowEasy does differently.

🔧 Auto-Heal AI

Traditional CI/CD tools send you a notification when a pipeline fails. That's it. You wake up, dig through logs, figure out the root cause, fix the YAML, push again, and pray.

FlowEasy's Auto-Heal AI takes a different approach:

  1. Reads the full error logs — not just the last line, but the entire context
  2. Identifies the root cause — using AI to understand what actually went wrong
  3. Commits the fix or opens a PR — automatically, without you lifting a finger

Average fix time: under 30 seconds.

Common issues it handles automatically:

  • Dependency version conflicts
  • Missing environment variables
  • YAML syntax errors
  • Test runner misconfigurations
  • Build cache invalidation issues

You set up the pipeline once. FlowEasy keeps it running.

🖥️ MCP Server — Manage Pipelines from Your IDE

One of the features I'm most excited about is the MCP (Model Context Protocol) Server. It gives you 9 tools you can use directly from Claude Code, Cursor, or Windsurf — without ever leaving your IDE.

Available tools:

  • check_pipeline_status — see what's running or failing
  • view_logs — pull full logs for any run
  • rerun_pipeline — trigger a new run
  • rollback_deployment — revert to the last working version
  • analyze_failure — ask AI to explain what went wrong
  • list_pipelines — see all configured pipelines
  • get_pipeline_config — inspect current configuration
  • update_pipeline — modify settings without leaving your editor
  • cancel_run — stop a running pipeline

One-line setup in your MCP config:

{
  "mcpServers": {
    "floweasy": {
      "command": "npx",
      "args": ["floweasy-mcp"],
      "env": {
        "FLOWEASY_API_KEY": "your-api-key"
      }
    }
  }
}

Works with Claude Code, Cursor, and Windsurf out of the box.

🛡️ 6 Security Scans — Zero Config, Security by Default

Every single deploy automatically runs 6 security scans. No extra configuration needed. No separate security pipeline to maintain.

Scan Tool What it catches
SAST Semgrep (847+ rules) Code vulnerabilities, injection flaws
SCA OWASP Dependency-Check Vulnerable dependencies
Secrets TruffleHog Exposed API keys, tokens, credentials
DAST OWASP ZAP Runtime vulnerabilities in running apps
SBOM Syft Full software bill of materials
Compliance Custom rules GDPR, SOC 2, policy violations

Security isn't bolted on as an afterthought. It's baked into every deploy from day one.

How It Works

Getting started takes about 5 minutes:

1. Connect your repo
Link your GitHub repository to FlowEasy. It analyzes your stack automatically.

2. Choose your gates
Select which security scans to enforce, what tests to run, and your deployment targets.

3. FlowEasy generates the YAML
No manual YAML writing. FlowEasy creates the GitHub Actions workflow for you, with all your selected security scans and Auto-Heal AI configured.

From there, every push triggers the pipeline. If something breaks, the AI fixes it. If a security scan fails, you get a detailed report with remediation guidance.

Stack

For those curious about how it's built:

  • Backend: Node.js + TypeScript
  • Pipeline Engine: GitHub Actions
  • AI: Claude API (Anthropic)
  • Security Tools: Semgrep, TruffleHog, OWASP ZAP, Syft
  • MCP Server: Model Context Protocol
  • Database: PostgreSQL

Try It

FlowEasy is live at floweasy.dev.

Free tier: 1 repo, 5 pipeline runs/month
Pro: $19/month, unlimited repos and runs

I built this to solve my own pain. If you've ever debugged a broken CI/CD pipeline at 2 AM, I think you'll find it useful.

What security scans matter most to you in your pipelines? SAST? Secrets detection? DAST? I'd love to hear what your current setup looks like and what gaps you're trying to fill.