petterjuan/agentic-reliability-framework

Performs multi-stage incident analysis through detection, diagnosis, and predictive reasoning on reliability events, then surfaces structured advisory remediation plans with business impact estimation—stopping at advisory intent without autonomous execution. Built on graph-native architecture for pattern recall across historical incidents, with deterministic safety guardrails enforced through configuration-driven policies to constrain recommendations scope and impact.

Available on PyPI.