Close the four review blockers on the merged run_stage extraction

[briansrls]

Follow-up to #7499 (merged as 7cff3d72b). These four changes were requested in the second-pass review but were not in the merged head, so they land as a fresh change on a branch restarted from main rather than stacked on merged history.

---

1. The canonical note contradicted itself

walk_plan_note asserted both that heavy whole-tree stage claims refuse — the current implementation — and that "the arm-time validator no longer refuses a heavy-whole-tree-resolve claim", which was stale text from the revision in between. Because this is the canonical carrier rather than incidental commentary, the generated stage0 projection reproduced the contradiction verbatim.

Replaced with a closed enumeration of all five refusals. The shape is the fix, not just the wording: a partial list written as commentary on what changed is exactly how a carrier ends up stating two incompatible things.

An on-success stage refuses, at arm time: an undeclared resource profile; heavy_whole_tree_resolve; spawns_host_compiler; substantial residency; a discovery runnable.

The dissolve-on was also wrong, and in a way that matters. The four profile restrictions name different triggers and do not dissolve together, and the undeclared-profile refusal has no dissolution at all — it is the fail-closed floor. The heavy trigger is two ordered steps: split execution-slot accounting from resident-reservation accounting, then take a context-lifetime reservation. Naming "a resident lease" alone understated it, because a lease against today's single active counter deadlocks: a resident hold pins active >= 1, so the active == 0 progress floor never fires, every later admission returns Hold(WindowFull), and the width that would relieve it grows only on a completion that can no longer happen. The runner starts at target_width=1, so that is the default path.

2. Rust comments still stated the retracted contract

Three phrases survived the carrier's retraction: "members run concurrently", "same governor admission", "same derived cost clamp". All now match what the executor actually provides:

Distinct spawned resolve groups may overlap, subject to governor admission. Same-entry groups, memo-lane units, main-thread units, and width-constrained spawned units may execute serially. No sibling order is guaranteed.

"Same clamp mechanism", not "same derived clamp" — ordinary batches supply clamp parameters; stages deliberately pass None and stay inside a narrow admissible profile instead.

3. Receipt identity was half closed

Per-stage receipts were attempt-scoped and payload-stamped. The aggregate had identity in neither path nor payload — while #7499's body claimed identity was closed before the first consumer.

Both aggregate write sites are now scoped and stamped, including the skip case ("the ordinary floor failed so no stage ran" is an answer about this attempt), which refuses rather than writing unattributably.

entry is now carried on ClaimResult rather than recovered by lookup. Searching a stage's runnables for a matching function name would reproduce exactly the ambiguity that makes a function name insufficient as a declaration identity. 16 construction sites. The two with no single declaring entry — the unmapped-node sentinel and the discovery aggregate — say so explicitly, because a blank field reads as "unknown" when the truth is "not one entry". plan_site added to both receipt headers, completing the subject location: attempt says which run, plan site says which plan, entry+function say which declaration.

4. The §7 deferral belongs at its own carrier

walk_plan_run_stage_claim_executor_seed_deferral is authored in std.realization_schedule, with the matching SCAFFOLD marker on the Rust it covers.

A first draft had it in the downstream fixture branch, on the reasoning that the fixture is where the seed expansion became visible. That reverses ownership: the debt belongs to the change that added the Rust, and a consumer documenting its parent's deferral means the parent can land without one. A consumer may cite the row; it may not be its home.

---

What this found about the gate itself

Four #[cfg(test)] sites were missing the new entry field. They were found by running the suite unfiltered — and they are worth calling out because of why nothing caught them earlier:

• cargo build does not compile test targets, so every "build clean" check in this arc was structurally incapable of seeing them.
• The verification chain piped the suite through grep -E "^test result:", so the compile error rendered as an empty section rather than as a failure.
• CI would not have caught them either: the build job compiles the binary, not test targets, and the executor suite was removed from CI in July. A compile error in test-only code passes full CI today. That is a standing gap, independent of this PR.

Both local gates are corrected: cargo test --no-run as the compile gate, and the suite piped through tail so failure reads as failure. This is the third instance of the same class in this arc — a check that cannot observe the thing it claims to check — after the exit-0-through-a-grep parse error and the dag/tools exhaustiveness miss.

Verification

• Test-binary compile, regen, fmt, and the full 48/48 suite were green on the pre-rebase tree carrying these exact changes.
• The same gate is re-running against this rebased base. CI on this head is the arbiter; if regen produces a seed update on the new base it will follow as a separate commit.

Not here

The admission occupants, and the ci_floor_declared_resolve_count raise they require, remain their own change — with the delta to be measured rather than assumed, since the carried "1→2" is analysis, not measurement.

🤖 Generated with Claude Code

https://claude.ai/code/session_01N7LfCpidMuDjV3FxYUHv1K

---

Generated by Claude Code

[cursor]

Bugbot is not enabled for your account, so this pull request was not reviewed.

Enable Bugbot in the Cursor dashboard to get automatic reviews on future PRs.

[gunbai-bot]

claim_executor test receipt (6fbbcc6f56)

Unfiltered merge-bar gates on the current head:

Test-binary compile (--no-run):

cargo test -p v1-compiler --bin claim_executor --no-run
    Finished `test` profile [unoptimized] target(s) in 3m 29s
  Executable unittests src/bin/claim_executor.rs (target/debug/deps/claim_executor-f00ba678dadb229a)

exit 0

Full suite (unfiltered, no grep):

cargo test -p v1-compiler --bin claim_executor
running 48 tests
…
test result: ok. 48 passed; 0 failed; 0 ignored; 0 measured; 0 filtered out; finished in 508.54s

exit 0

— sent from silent-tern-582 (dispatch A, running #7518 bar on shared worktree)