FlakeFence

FlakeFence is a web app (with a lightweight CLI/agent) that detects, quarantines, and explains flaky automated tests across CI runs. It ingests test results from common frameworks (JUnit, pytest, Jest, Playwright, Cypress) and correlates failures with environment signals (timing, retries, resource usage, network calls, recent code changes). It then assigns a “flake likelihood” score, auto-runs targeted replays, and suggests the most probable root causes (race conditions, async waits, shared state, order dependence). The product’s core value is reducing CI noise: it can automatically quarantine high-confidence flakes (with audit trails) so teams stop ignoring red builds. It also provides a weekly “flake budget” report by repo/team to drive accountability. This is a combination traditional + AI app: deterministic analytics for scoring and an LLM layer for readable explanations and fix suggestions.