diff --git a/plans/abstractions.md b/plans/abstractions.md
index a8fa4a1e..72968bd7 100644
--- a/plans/abstractions.md
+++ b/plans/abstractions.md
@@ -12,6 +12,11 @@ Anything short → _Watching_ (what's missing) or _Rejected_ (why).
 
 ## Last scan
 
+- **Date:** 2026-06-07 (radar loop, pass 29)
+- **Pass 29:** conformance loop excluded `js` (test262 fixtures) → 4 migrated + 2 excluded,
+  3 remain (ocaml/smalltalk/tcl). New subsystems advancing fast: `relations` → Phase 4
+  federation, `artdag` → Phase 6 federation → both fold into W1 (now 7 federation modules,
+  theme-not-shape holds) and W9 (relations past Phase 2 but not yet consumed by anyone).
 - **Date:** 2026-06-07 (radar loop, pass 28)
 - **Pass 28 — fleet expanding again.** Conformance loop: `go` migrated 609/609; **`forth`
   excluded** (foreign Forth corpus — classify-then-exclude working). 4 migrated +1 excluded
@@ -251,8 +256,9 @@ one merged file copied N times. Correct one-liner:
   prolog, search — `events` migrated via its OWN loop; `search` via counters mode (which
   corrects the earlier "search excluded" note). **+4 on the `loops/conformance` branch:
   `common-lisp` 487/487, `erlang` 761/761, `feed` 189/189, `go` 609/609** — pending merge.
-  **`forth` EXCLUDED** (foreign Forth corpus — Hayes core.fr via awk+python, like lua).
-  Remaining for the loop: js, ocaml, smalltalk(likely-excluded), tcl.
+  **`forth` + `js` EXCLUDED** (foreign corpora: forth=Hayes core.fr via awk+python;
+  js=test262 `.js`/`.expected` fixtures — neither is SX suites, like lua).
+  Remaining for the loop: ocaml, smalltalk(likely-excluded), tcl.
 - **NOW IN PROGRESS — dedicated loop (2026-06-07).** A human-triggered `conformance` loop
   (worktree `/root/rose-ash-loops/conformance`, branch `loops/conformance`, tmux session
   `a1-conformance`, briefing `plans/agent-briefings/conformance-loop.md`) is working the
@@ -306,6 +312,11 @@ one merged file copied N times. Correct one-liner:
   real transport, these 4 become its *consumers* (wiring `send-fn`/`fetch-fn`/`transport`
   to it) — that work belongs to each subsystem's loop + the `fed-sx` loop, not a
   cross-cutting extraction. Stop re-proposing on the shared name. Home: `fed-sx`.
+- **Now 7 federation modules (pass 29):** + `relations` (Phase 4: erel trust-gating,
+  peer_rel/trust, fed-sx mock transport — Datalog-rule trust like acl) and `artdag`
+  (Phase 6: content-addressed cache + trust + **invalidation** — a merge shape unlike any
+  other). Each new one reinforces "theme not shape": 7 divergent merges, all sharing only
+  the inject-fed-sx-transport seam. Verdict unchanged — they're fed-sx consumers-in-waiting.
 - **Narrower sub-claim (mod note, pass 6; refined pass 18):** mod asserts the *fed
   trust/outbox* shape shares between mod+acl. Radar evidence refines this: the trust gate
   splits by mechanism, not by subsystem pair — **mod + events** both use a runtime
@@ -444,6 +455,11 @@ one merged file copied N times. Correct one-liner:
   local impl is currently domain-specific (different keys/semantics). Do NOT propose yet.
   Re-check when relations is past Phase 2 AND ≥3 subsystems' relationship logic could
   genuinely delegate to it. `artdag` also just spawned (nascent, 0 files) — tracking only.
+- **Update pass 29:** relations rocketed to **Phase 4** (one gate — past Phase 2 — now met),
+  but it's building ITSELF out (schema/federation), **not yet being consumed** by anyone.
+  The blocker is the other gate: 0 subsystems currently *delegate* their parent/child logic
+  to it (feed/content/events still track locally). Watch for the first real delegation.
+  (artdag also raced to Phase 6 — these ports advance fast; treat committed state as truth.)
 
 ### W8 · Durable externally-resumed orchestration on `lib/flow` (suspend→host-IO→resume)
 - **The shared shape:** a durable `flow` that `request`s an external action (a suspend