Forge STI

Software Testing Intelligence for governed delivery. STI structures test strategy, plans, cases, and campaigns as reviewable artifacts — authored in Cursor or by agents, orchestrated with lmeta, reasoned with LCDL where…

Updated

What is STI? · Dual-wiki architecture · Schemas

Choose your path

Trust and boundaries

  • STI does not replace pytest or Playwright — it governs how test intent is authored, linked, and evidenced.
  • LCDL owns judgment — case design critique, triage, and coverage analysis use governed tasks when they affect decisions.
  • Fleet owns execution — containerized test runs attach logs to EvidencePacket; STI does not run Docker.
  • Lenses owns review — test evidence surfaces for human approval; STI does not bypass the control plane.