tracer-cloud/opensre — reverse-engineered prompt

Build me OpenSRE, an open source Python tool for teams that get production alerts and need an AI helper to investigate what went wrong.

I want someone to install it, run an onboarding command, connect their LLM provider and monitoring tools, then either chat with it in a terminal or give it an alert file. It should pull together alert details, logs, metrics, traces, runbooks, and related context, reason through the incident, and return a clear report with the likely root cause, evidence, and next steps. If configured, it should be able to post a summary to Slack or PagerDuty and support safe optional remediation actions.

Please include the basic command line experience, a deployable service version, environment based configuration, Docker support, and tests for synthetic incidents and real infrastructure style scenarios. Keep it usable as a public alpha, with simple docs for install, quick start, deployment, integrations, and security. Look up current docs online if you need to.