Classification: layer=H (weak-reader) + A (language-shape); mechanism=scaffold I.

Intent: Offer a capability-demanding multi-file change as ONE parameterized, checked operation — converting an edit a weak agent cannot safely coordinate into a parameter-filling task. TypeScript's mature codemod ecosystem (`ts-morph`, `jscodeshift`, typed ESLint autofix) makes this the most achievable scaffold here, where in Rust it is the least.

Also Known As: codemod; AST rewrite; refactoring script; scripted migration; semantic patch; `jscodeshift` transform; ESLint autofix.

Applicability / Recognition: Apply when — a common change touches many files atomically (add a cell, register a variant, rename across a seam); the edit's size is itself the failure driver (R2C-006); the weakest swarm tier cannot coordinate it by hand. *Detector seed:* a recurring change-type that reliably requires touching >1 file in lockstep → recognition fires.

Motivation: A weak agent asked to "rename this seam across its 7 call-sites + the barrel re-export + the discriminated error union" desynchronizes them. A `codemod rename-seam --from X --to Y` over TypeScript's mature AST tooling **would** perform the change atomically and verifiably, the agent filling two parameters instead of coordinating seven edits — that operation is specified and not yet built. The shipped codemod surface today is one verb, `typescript-ai-native codemod add-cell --cell <cell> --spec-uri <uri>`, which writes the seam module and its `node:test` smoke test atomically and rolls back on failure. This mirrors how constrained decoding lifts weak models (DR1-015): collapse the hard task into a constrained, parameterized one.

Structure & Participants: *Codemod* (`ts-morph`/`jscodeshift` AST rewrite, or typed ESLint autofix) · *Parameters* (the small named inputs) · *Atomic application* (all-or-nothing) · *Post-check* (`tsc` + `vitest` green).

Collaborations: Implements bulk application of Classes A/B/G in raids; emits Class F diagnostics on failure; the Class D oracle wraps it when it changes behavior.

Goals / Non-Goals: *Goals:* convert capability-demanding multi-file edits into parameterized operations for the weak swarm. *Non-Goals:* NOT a general refactoring IDE; NOT for one-off changes; NOT a production transform of semantics.

Consequences: (+) the weakest tier can perform edits otherwise beyond it; (+) atomicity kills desync and phantom diffs; (+) the tooling is mature, so the build-side cost is low — TypeScript's scaffold advantage. (−) codemods are code to maintain and test; (−) **[E-hyp] risk:** parameterizing a codemod may itself exceed the weakest models — the very build/use boundary in question.

Alternatives: hand-editing (fails at scale for weak agents); a generator (Class A) when the artifact is derivable rather than transformed. Codemods are for TRANSFORMING existing code.

Risks & Assumptions: **assumes weak agents can correctly parameterize the operation** — UNVALIDATED; this is the prime pilot question. If false, restrict the weakest tier to fixed-parameter invocations only. Unlike Rust, the tooling-immaturity risk does NOT apply here — `ts-morph`/`jscodeshift` are production-grade; only the parameterization question remains open. *Sunset:* if language/tooling makes the change trivial, the codemod retires.

Evidence & Transfer-strength: first-principles from R3-013 (ownership graph bounds throughput) + R2C-006 (edit size drives failure) + DR1-015 (constraints lift weak models). The mature ecosystem is the one place TypeScript moves the [E-hyp] tag toward feasibility (the *build* half is solved; the *use*/parameterization half is the open question). Class: theory. Tag: **[E-hyp]**.