Classification: layer=E (verification) + H (weak-reader); mechanism=scaffold E.

Intent: Guarantee every cell is independently type-checkable and testable in seconds (`tsc --noEmit -p <cell>` + `vitest run <cell>`), so an agent's edit→check→read-error→edit loop gets a first signal fast enough to steer — the substrate that makes every other scaffold's pass/fail usable.

Also Known As: tight feedback loop; incremental check; project-reference isolation; per-project typecheck; watch mode; agent-computer interface.

Applicability / Recognition: Apply when — a cell cannot be checked without type-checking/building the whole repo; the only verification is a multi-minute CI run; an agent must wait minutes for a signal. *Detector seed:* `tsc --noEmit -p <cell>` or `vitest run <cell>` fails to run in isolation, OR the cell has no project reference / vitest project → recognition fires (verification locality beats capability, R3-007).

Motivation: A weak agent in a bounded loop gets one signal per multi-minute CI run — useless. With `tsc --noEmit -p <cell>` + `vitest run <cell>` returning in seconds (sub-second under the TS 7 / "Corsa" native compiler, ~10× faster checking), the same agent iterates ten times in the budget that previously bought zero feedback. The interpreter-budget result confirms: more local runs help agents that can use feedback; the loop is the amplifier.

Structure & Participants: *Cell isolation* (a project reference / a `vitest` project) · *Fast checker* (`tsc --noEmit -p` + `vitest`) · *Structured error* (Class F) · *Budget guard* (first signal < ~60s).

Collaborations: Runs Classes C/D/G checks and the type-level tests (guide §12); consumes Class B's compile checks; emits Class F diagnostics. The raid executor runs this per batch.

Goals / Non-Goals: *Goals:* sub-minute first signal per cell; make the loop the standard agent workflow. *Non-Goals:* NOT replacing full CI (still runs at merge); does NOT create capability — scaffolds amplify, they do not create.

Consequences: (+) every other scaffold becomes loop-usable; (+) iteration count rises within budget; the native Corsa compiler sharpens this further. (−) requires cells genuinely isolable (drives §3 cell design); barrel-file sprawl breaks project-reference isolation; (−) fast checks may miss what full CI catches — gate-mode checks backstop.

Alternatives: whole-repo `tsc`/CI only (too slow for an agent loop); manual testing (not mechanical). Neither is a substitute.

Risks & Assumptions: assumes cells are isolable; god-modules and giant barrels break this (anti-pattern). *Sunset:* none — the loop is foundational.

Evidence & Transfer-strength: R3-007 (verification locality, theory), R2C interpreter-budget result (more runs amplify capable agents, benchmark), BLD-010 (30–60s practical threshold). Class: benchmark + theory. Tag: **[E-strong]**.