What problem does it solve? Running every test suite, bench, and gate in a large Node.js framework repository in the wrong order produces false failures, misleading green totals, and verdicts that confuse saturation or missing infrastructure with real product regressions. ## Core Features & Use Cases - Ordered six-stage campaign matrix: Sequences non-regression, load and memory, cost switches, load benches, scaffold verification, and multi-pod tests so each stage does not corrupt the next. - Red-verdict decision tree: Instructs each failure through isolation replay, decor checks, shared-state accumulation, instrument staleness, and only lastly the product code. - Blind-spot accounting: Tracks closed switches, skipped benches, out-of-scope scripts, and the unpublished release chain so a green total is never reported without what did not run. - Use Case: Before publishing Nodefony 10.0.0, run the full campaign stage by stage, triage every red by cause, and generate the committed HTML quality report under docs/qualite. ## Quick Start Ask the agent to run a complete Nodefony test campaign and produce the final HTML report with every red verdict instructed by cause.
