docs: clarify that skipped/#240 shipped in PR #249, not this PR

The proposal previously framed the #240 skipped-node work as part of
'this PR'. That work landed earlier in PR #249; this PR (#263) defines
the consolidated state model and adds the remaining implementation
(cancelled, status_reason, race-loser reconciliation, loop reset).

Author: pinodeca

docs/node-state-model.md

@@ -1,4 +1,4 @@
-# Node & Instance State Model
+occ# Node & Instance State Model
 
 Status: **Design proposal** (consolidates issues [#240] and [#171]).
 
@@ -93,10 +93,10 @@ Rationale:
 | `skipped` | yes | Eligible work that was never started because the run already terminated, or a branch the run chose not to take. **(#240, new in 0.2.4)** |
 | `cancelled` | yes | Work that had started (`running`) or was queued and was abandoned because its enclosing scope (a decided `RACE`) no longer needs it. **(#171, proposed)** |
 
-> Current state of the code: `pending`, `running`, `completed`, `failed` exist;
-> `skipped` is added by this PR (issue #240). `cancelled` for **nodes** is proposed
-> here for #171 (the `df.instances` table already has a `cancelled` status, but
-> `df.nodes` does not yet).
+> Current state of the code: `pending`, `running`, `completed`, `failed`, and
+> `skipped` exist; `skipped` was added in 0.2.4 for issue #240 (PR #249).
+> `cancelled` for **nodes** is proposed here for #171 (the `df.instances` table
+> already has a `cancelled` status, but `df.nodes` does not yet).
 
 ### `status_reason` vocabulary (proposed)
 
@@ -337,10 +337,10 @@ before committing to the reset mechanism.
 ### #240 — downstream steps after a failure
 
 When a run terminates with a node-level failure, remaining un-started nodes should
-become `skipped` with reason `upstream_failed`. The current PR implements the
-set-based sweep (`pending → skipped` guarded by "an instance node is `failed`").
-Under this model that sweep should additionally stamp `status_reason =
-'upstream_failed'`.
+become `skipped` with reason `upstream_failed`. The `skipped` status (#240, shipped
+in PR #249) already implements the set-based sweep (`pending → skipped` guarded by
+"an instance node is `failed`"). Under this model that sweep should additionally
+stamp `status_reason = 'upstream_failed'`.
 
 Open refinement: the coarse sweep also catches `IF`/`LOOP` branches that were
 *deliberately* not taken. With the reason column those should ideally be classified
@@ -370,9 +370,9 @@ Only the final iteration's loser marks survive as the observable result.
 
 ## Schema & implementation implications
 
-- **Status set:** add `skipped` (done, #240) and later `cancelled` to
-  `nodes_status_chk` in both the install DDL ([src/lib.rs](../src/lib.rs)) and a new
-  upgrade script, per the upgrade contract.
+- **Status set:** `skipped` is already added (done in PR #249 for #240); add
+  `cancelled` to `nodes_status_chk` in both the install DDL
+  ([src/lib.rs](../src/lib.rs)) and a new upgrade script, per the upgrade contract.
 - **Reason column:** add `status_reason TEXT` to `df.nodes` (nullable, optionally a
   `CHECK` against the reason vocabulary). New column ⇒ install DDL + upgrade DDL;
   binary backward-compat is straightforward (new `.so` only writes it when the