Debug Vitals

Developers increasingly use multiple AI coding agents to write code, fix bugs, install dependencies, and run commands. This creates "fragmented context". One agent changes the code, another changes a dependency, Git records another change, while CI reports the final failure separately. When something breaks, developers have to manually reconstruct what happened: "What did the agent change?", "Which dependency changed?", "Did the environment differ?", and "What actually contributed to the failure?" **Debug Vitals connects the dots between development changes and failures.** It builds a traceable chain: **Agent Action → Code/Dependency Change → Environment → CI Failure** Instead of only explaining an error from a log, Debug Vitals uses evidence across these stages to help developers understand "what changed" and "what is connected to the failure." The core capabilities include "Agent Activity" to record meaningful agent actions, "Dependency Intelligence" to track dependency and lockfile changes, "Environment Comparison" to identify relevant runtime and tool differences, and "CI Failure Correlation" to connect changes with failures. It also provides "Trust Policies" that can flag sensitive or destructive actions for approval, along with "Evidence-Based Findings" that separate observed evidence from likely relationships instead of blindly guessing a root cause. Debug Vitals is built for developers and teams using AI coding agents, Git-based workflows, and CI pipelines who want to spend less time reconstructing failures and more time fixing them. What makes it different is its focus on the "chain of events behind a failure". It connects what an AI agent did with what changed in the project and what eventually failed, rather than simply presenting another log or monitoring dashboard. Debug Vitals is already packaged and published on npm as **"debugvitals"** and can be run directly with: `npx debugvitals` The project can evolve into IDE extensions, MCP integrations, and language-agnostic tooling, allowing the same debugging intelligence to work across different agents and development environments.