@@ -114,6 +114,16 @@ task touches:
114114- Ref generation pin: optional ` ~s<n> ` suffix on an @ref carrying the snapshot generation it was
115115 minted from. Accepted as input everywhere, emitted by no tree output (snapshot token budget),
116116 auto-appended by the MCP layer, stripped and ignored by replay.
117+ - Deferred interaction outcome: the daemon's post-response answer to "did that mutation actually
118+ take effect" — pending interaction outcome retry, post-gesture stabilization, and Android
119+ snapshot freshness recovery. ` src/daemon/deferred-interaction-outcome.ts ` is its one interface:
120+ every mutating route marks through it after dispatch, and every snapshot capture resolves
121+ through it; the three ` SessionState ` fields stay with their R7 owners (the module itself owns
122+ ` postGestureStabilization ` , so the seam adds no node to the R9 cycle). Marking
123+ order is load-bearing (pending outcome retry before stabilization), each marker keeps its own
124+ eligibility gate, and the module owns only these post-action markers — never ADR 0014 ref-frame
125+ expiry or the ADR 0012/0016 staged protocols. Distinct from the same-response settled
126+ observation below.
117127- Settled observation: opt-in (` --settle ` ) post-action payload on press/click/fill/longpress — the
118128 quiet-window stable loop re-captures until the UI settles, and the response carries the diff vs the
119129 pre-action tree (changed lines only, added lines with fresh refs, ` refsGeneration ` when the settled
@@ -312,16 +322,20 @@ The perfect-shape refactor is complete and merged. Its end-state:
312322 theirs.
313323- Type-cycle growth (R9). R4 keeps the VALUE import graph acyclic, so every remaining cycle is
314324 created by type-only imports — free at runtime, invisible to R5/R6, and the largest single
315- obstacle to reading a subsystem in isolation: inside a strongly-connected component of 76 files,
325+ obstacle to reading a subsystem in isolation: inside a strongly-connected component of 49 files,
316326 no file has a self-contained slice. ` TYPE_CYCLE_BASELINE ` , derived from the zone ceilings in
317327 ` scripts/layering/daemon-modularity.ts ` , ratchets it for ** growth only** , deliberately unlike R6: reducing it
318328 is a real refactor rather than a file move, so a hard equality would turn every unrelated
319329 improvement into a baseline edit. A shrunk tree is reported in the success line instead of
320- failing. Hubs by in-component dependents: ` runtime-contract.ts ` (25),
321- ` commands/runtime-types.ts ` (21), ` backend.ts ` (15), ` commands/runtime-common.ts ` (12).
330+ failing. Hubs by in-component dependents: ` core/dispatch.ts ` (8),
331+ ` command-catalog.ts ` (7), ` commands/interaction/runtime/resolution.ts ` (6),
332+ ` core/command-descriptor/registry.ts ` (6). The former type hubs (` runtime-contract.ts ` ,
333+ ` commands/runtime-types.ts ` , ` backend.ts ` , ` commands/runtime-common.ts ` ) left the cycle when
334+ #1632 sank ` backend.ts ` 's two upward type imports — 27 files stranded out of the component at
335+ once.
322336- Daemon modularity ratchets (R10). The same tooling-only declaration pins R7's writer-owned
323- field/owner-claim counts, R9's 76 members by zone (` commands ` 33 , ` daemon-server ` 20 , ` core ` 10,
324- ` platforms ` 7 , root 5 , ` client ` 1), and the external production importers of ` daemon/types.ts `
337+ field/owner-claim counts, R9's 49 members by zone (` commands ` 14 , ` daemon-server ` 19 , ` core ` 10,
338+ ` platforms ` 2 , root 3 , ` client ` 1), and the external production importers of ` daemon/types.ts `
325339 (down to 2: the client normalizers and remote artifacts). R7 counts and external importers may
326340 only shrink; no zone may grow inside R9, and replay/Maestro/replay-test engine files remain outside
327341 it. This per-zone ratchet is intentionally stricter than R9's ordinary total-growth rule: even
0 commit comments