Developer builds local AI incident commander tool, claims 88% pass rate across 15 outages

A developer has released ai-incident-commander, an open-source CLI tool designed to handle documentation tasks during IT incidents, including timelines, stakeholder updates, and postmortem drafts. The tool runs entirely on a local machine using a local LLM, requiring no API keys or cloud services, and is available under an MIT license via GitHub. Built after the developer's frustration with incident commanders being overwhelmed by simultaneous documentation and decision-making duties, the tool includes three human-approval checkpoints and never executes any production actions. Testing against 15 real-world outage postmortems from companies such as Cloudflare, GitLab, and AWS yielded 106 out of 120 checks passed, an 88.3% pass rate, at a total LLM cost of approximately $0.02. The developer noted that early test failures stemmed from flawed evaluation criteria rather than tool shortcomings, and adjusted the benchmarks to measure coherence and quality over exact text matching.