• ROOT CAUSE
  • LOG SUMMARY
  • REPAIR PLAN
  • HUMAN REVIEW

How to debug failed builds with your AI agent

Connect Concord to your agent to stop digging through raw text. Once connected, your agent finds the exact process that failed and reads the logs for you.

Ask AI about this page

Short answer

How can my AI help me debug a failed build?

Your agent uses Concord to find the specific failed process and pulls its logs to pinpoint the error. You get a clear explanation of the failure without manual searching. This breakdown shows you the exact steps it takes.

The results of your analysis.

Where your debugging efforts land.

The agent sorts its findings into four distinct categories based on what it finds in the logs.

ROOT CAUSE

Identified error source

The agent finds the specific process that triggered the failure. You get the exact error message and the context surrounding it.

LOG SUMMARY

Condensed execution history

The agent filters out the noise from the execution stream. You get a summary of the relevant events instead of thousands of lines of text.

REPAIR PLAN

Suggested fixes

The agent compares the error to the log history to suggest what to change. It provides a starting point for your fix.

HUMAN REVIEW

Manual intervention needed

If the logs are inconclusive or the error is an environmental issue, the agent flags it for you. You decide how to proceed.

The workflow

What your AI does when a build fails.

The agent moves from broad process discovery to deep log inspection to find the problem.

  1. Locate failed processes

    The agent scans your recent Concord runs to find which specific tasks or processes ended in a failure state.

    list_processes

  2. Extract error details

    Once a failed process is identified, the agent pulls the specific output and error streams for that execution.

    get_process_log

  3. Analyze log patterns

    The agent reads through the retrieved logs to find stack traces, timeout errors, or permission issues.

    get_process_log

  4. Report the cause

    The agent summarizes the findings and presents the most likely reason for the deployment crash.

    get_process_log

Try it

Copy these to start.

Use these prompts to kick off the debugging process.

Starting points

Swap the bracketed text for your actual deployment names or process IDs.

Concord Connector

You're all set. Choose your MCP client and follow the setup instructions.

Connector linkhttps://edge.vinkius.com/vk_preview_Zu1FtaypR6vodFhZs1KNTeO3BBzUOEVSxrwO3xgK/mcp

Claude Desktop

Follow the steps below to connect in seconds.

  1. 1In Claude Desktop, open Settings → Connectors.
  2. 2Click “Add custom connector” and paste the connector link above as the remote MCP server URL.
  3. 3Click Add and start a new chat — Concord capabilities are ready to use.
Configuration · claude_desktop_config.jsonCopy
{
  "mcpServers": {
    "concord-workflow-orchestration-mcp": {
      "url": "https://edge.vinkius.com/vk_preview_Zu1FtaypR6vodFhZs1KNTeO3BBzUOEVSxrwO3xgK/mcp"
    }
  }
}
Copy into chat04

  • List all the processes from my last Concord deployment so I can see which one failed.

  • Get the logs for process 'deploy-api-service' and tell me why it crashed.

  • Look at the logs for the failed 'build-frontend' process and find the specific error message.

  • Check the logs for the most recent failed execution and summarize the root cause.

  • Claude
  • ChatGPT
  • Cursor
  • VS Code
  • Windsurf
  • Claude Code
  • JetBrains
  • Cline

Start here

Connect Concord once, then ask.

Just link your account once. Your credentials stay encrypted, and you can immediately start asking your agent about your Concord workflows.

Connect Concord to your AI

FAQ

How this task behaves

  • 01

    Can the AI fix the build automatically?

    No. The AI only reads logs and identifies the cause. You have to apply the fix yourself.

  • 02

    Can the AI delete failed deployment processes?

    No. The tools only allow the agent to list processes and read their logs.

  • 03

    Does the AI need my Concord password?

    No. You connect via Vinkius, which manages the connection securely.

  • 04

    Can the AI see all my historical data?

    The AI can only see the data it can access through the list and log tools provided by the MCP.

  • 05

    What if the logs are too large?

    The agent uses the log tool to pull the necessary information to find the error, focusing on the relevant segments.

  • More questions about Concord? The Connector page answers them. See everything the Concord Connector can do

Connect Concord to Claude, Cursor, ChatGPT & more