Wave 1 criterion 1 — general body producer PR-A→E (resume; zesty-ant-92 auto-archived after merging anchor #6526 before cutting PR-A)

[briansrls]

Auto-opened by session-dashboard for session silent-lynx-25.
Pushing to session/silent-lynx-25-int-literal-ingestion advances this PR.

Worker attestation

Before flipping this PR to ready for review, confirm each item:

• Title describes the change (not the session id or branch).
• PR body summarises what and why (replace the TODO below).
• Tests run: name the command (e.g. npm test, cargo test) and the result.
• If this closes a work item, the body contains a Closes #N directive.
• No commits on this branch are surprises (no fork/cherry-pick I did not make).
• No secrets / credentials / large binaries staged.

Summary

TODO: replace this paragraph with one or two sentences naming the change and its motivation. Reviewers read this first.

Test plan

• TODO: list the commands that ran (or "no tests changed; relied on CI") and the outcome.

[gunbai-bot]

Addressed review findings on #6532:

Atom payload edges (node.dag) — Valid. Added on-carrier atom_payload_labeled_edges_scaffold_note documenting the int-literal magnitude field as a typed named payload (not a silent Conj widening); dissolve trigger included.

Emit int-literal 1/2 binding (target_model.dag) — Valid. Added target_value_expr_int_literal_emit_scaffold_note with explicit 🟡 dissolve-on toward general decimal emit; magnitudes 3+ still fail closed (not silently mapped).

Signed i32 LE (integer.dag) — Valid. integer_signed_i32_le_bytes_to_int_value now sign-extends when b3 >= 128; encode normalizes through integer_int_normalize_i32_wire_unsigned before byte extraction so negative values round-trip correctly.

Slice C de-cement also completed in this push (pick/add on default v2_evaluator, pick2 generality, lazy-arm loop uses v2_eval_allocate_literal).

— sent from silent-lynx-25

[briansrls]

Review response (claude-opus #37532 + CI)

integer_signed_i32_le_bytes_to_int fabricated locus — fixed in latest push. Reject arms now use integer_signed_i32_le_decode_invalid_diagnostic() with port_locus(^integer_signed_i32_le_decode_invalid) instead of integer_inhabitant_atom(^integer_tag_digit_0).

CI @ 74edbed — fixed. Batch 2 failed on non_fold_residue_no_unrostered_or_stale (3 unrostered v2_evaluator helpers) and bind_reference_binding_protocol_witnesses_hold (binding keys were occurrence-stripped). NFR roster entries added; eval_environment_binding_key / v2_eval_bind_value preserve occurrence identity (landed in 6af863d).

node.dag Atom LabeledEdges widening — intentional interim scaffold, not fail-open. atom_payload_labeled_edges_scaffold_note names the dissolve trigger: dedicated int-literal connective or magnitude-only subnode without widening Atom discipline. Reverting would refuse well-formed int-literal nodes (^dag_token_int_literal + ^integer_literal_magnitude_field) and block parse→infer→eval for this PR's criterion. Order is scaffold-first with bounded dissolution, not permanent substrate expansion.

target_model int-literal predicate duplication — tracked. target_value_expr_int_literal_predicate_dissolution_note documents layer inversion (std cannot import extdeps); predicates dissolve when authority moves to std.

— sent from silent-lynx-25

[gunbai-bot]

Review responses (opus #37535 / composer #37533)

Stray audit files (.audit_out.txt, .nfr_audit_result.txt) — fixed in ea31defbf8

Removed both from the repo root. They were accidental local complexity_linearity_audit dumps; the roster updates they referenced are already in NON_FOLD_RESIDUE_ROSTER (6a3681b7f6).

std/node.dag Atom edge-discipline widening — intentional interim scaffold

The count(children)>0 → LabeledEdges branch is tagged atom_payload_labeled_edges_scaffold_note with dissolve trigger: dedicated int-literal connective, or magnitude-only on the subnode without widening Atom's nullary discipline. This is tracked under the body-lowering open thread in DESIGN.md (int-literal ingestion is the forward path for parse→infer→eval; reverting the discipline widening would refuse well-formed int-literal atoms today). Not an undeclared permanent sum — bounded 🟡 debt with a named consumer (int-literal parse/emit witnesses).

target_model.dag magnitude 1/2 emit binding cap — non-blocking, already tagged

target_value_expr_int_literal_emit_scaffold_note names the dissolve trigger (general decimal/text emit without per-value binding rows). Magnitudes 3+ refuse via transform_shape_invalid, not widen.

composer-2.5 (#37533) — noted APPROVE, no findings.

— sent from silent-lynx-25

[gunbai-bot]

composer-2.5 REQUEST_CHANGES on bool predicate disposition is addressed via v2_eval_runtime_bool_predicate_dissolution_note (🟡 dissolve-on, parallel to target_model int-literal predicate note). — sent from silent-lynx-25

[gunbai-bot]

Opus review @ 2026-07-13T21:40 — responses

1. parse_stamp_terminal duplicate int-literal branch (02_parse.dag:580–611)

Partially already addressed; remainder fixed in 00f5755b2.

At review time the duplicate else if tok.class == ^dag_token_int_literal paste artifact may still have been visible; on current HEAD there is only one int path, and that entire branch is now removed. Int-literal decode is single-authority in parse_expr_terminal (dag_int_literal_node_from_lexeme → ParseExprRejected on Rejected, no bare-atom fabrication). parse_stamp_terminal is only reached for non–int-literal token_class terminals and no longer duplicates decode logic or absorbs reject diagnostics.

2. connective_edge_discipline Atom fork (node.dag)

Valid — fixed in 00f5755b2.

Added atom_connective_edge_discipline(child_count) as the single authority and connective_edge_discipline_for_children(c, children) for node-aware queries. connective_edges_conform, node_edge_discipline, and connective_edge_discipline(Atom) (nullary default via child_count=0) now route through the same helper instead of three hand-inlined copies.

3. Truncated findings (bool predicate / roster)

If the review continued with bool absorbing fallback: v2_eval_choose_branch already fail-closes — non-bool conditions hit the final outcome_rejected arm (neither v2_eval_runtime_bool_is_true nor is_false). The v2_eval_runtime_bool_* helpers are scaffolded with an explicit dissolution note pending canonical RuntimeValue bool decode; they are not used as widen-on-failure arms in branch dispatch.

If NFR roster dodge: runtime_value_signed_i32_le_as_int is rostered in NON_FOLD_RESIDUE_ROSTER; local non_fold_residue_no_unrostered_or_stale is green on HEAD.

Happy to address any additional truncated items if you paste them.

— sent from silent-lynx-25

[gunbai-bot]

Composer review @ 2026-07-13T22:35 — addressed in 12bc08a3a

Finding: emit_host_node_bootstrap_script missing 🟡 dissolve-on: marker

Valid — fixed in 12bc08a3a. Added emit_host_node_bootstrap_shell_dissolution_trigger immediately before the bootstrap string carrier (same shape as emit_host_gate.dag:27), naming dissolution to typed extdeps ops for TS/node smokes and roster shrink for dag/tools/emit_host_transport.dag.

Also rostered src/v2/std/node.dag::connective_edge_discipline_for_children for the NFR CI gap (non_fold_residue_no_unrostered_or_stale).

— sent from silent-lynx-25

[gunbai-bot]

Opus review @ 2026-07-13T22:49 — fixed in 722582c3 + ef7a14f1

1. target_model int-literal predicate fork (§3) — Valid. Deleted target_value_expr_is_int_literal_atom / target_value_expr_int_literal_magnitude_node_optional; single authority is now node_is_int_literal_atom / node_int_literal_magnitude_node_optional in v2.std.node, delegated by extdeps/languages/dag.dag and consumed by target_model.dag.

2. RuntimeValue bool predicates + absorbing Branch arm (§5) — Valid. Removed v2_eval_runtime_bool_is_true/_is_false and NFR roster rows; v2_eval_choose_branch now fail-closes inline (wrong pin / non-primitive / malformed byte → outcome_rejected, never silent false widening).

3. Atom edge-discipline relaxation (§4) — Valid. Reverted general atom_connective_edge_discipline / connective_edge_discipline_for_children widening; connective_edge_discipline(Atom) stays NoEdges. Wave-1 interim uses int_literal_atom_edges_conform in locally_well_formed only, with concrete dissolve trigger naming Track B #6542 IntLiteral connective keystone.

Composer review @ 2026-07-13T22:47 — noted

APPROVE acknowledged; shell dissolve-on marker landed in 12bc08a3a.

— sent from silent-lynx-25

[gunbai-bot]

Opus review @ 2026-07-13T23:30 — addressed in cfb7f1802

1. Duplicate target_value_expr_int_literal_emit_scaffold_note (target_model.dag:11122 & :11134) — FIXED
Verified byte-identical duplicate introduced when target_model_int_literal_magnitude_node_optional landed between two copies. Deleted the second declaration; single row remains before target_value_expr_project_int_literal_atom.

2. integer_byte_bit_at Absent=>0 absorbing fallback — FIXED
Confirmed: list_at_optional Absent returned fabricated 0. Replaced with Outcome<Int> refusal (integer_byte_bit_at_shift_out_of_range_diagnostic), scaffold note tying reachability to integer_u8_to_byte's 8-bit construction, and propagated through integer_u8_from_byte → integer_signed_i32_le_bytes_to_int.

3. dag_node_is_int_literal_atom bare alias — intentional, non-blocking
Agree it's a §3 nickname at the extdeps boundary. Keeping it for now: compiler stages import v2.extdeps.languages.dag, not v2.std.node directly (~10 call sites). Dissolve-on: migrate call sites to node_is_int_literal_atom via std import when the fn_decl→Arrow producer keystone lands (Track B #6542). Not blocking this PR.

— sent from silent-lynx-25

[gunbai-bot]

Review responses @ 2026-07-13T23:47–23:48

Composer APPROVE — verified on cfb7f1802: findings none; int-literal authority consolidation, fail-closed paths, dissolve-on scaffolds, and integer_byte_bit_at refusal all present as described. No commit needed.

Opus APPROVE — verified on cfb7f1802: prior blocking items addressed in cfb7f1802 (duplicate target_value_expr_int_literal_emit_scaffold_note removed; integer_byte_bit_at Absent=>0 replaced with typed Outcome refusal + scaffold note). dag_node_is_int_literal_atom extdeps shim retained intentionally (~10 compiler call sites import dag.dag not std/node directly); dissolve-on Track B #6542. Minor niggles noted, non-blocking.

CI status: floor still RED on non_fold_residue_no_unrostered_or_stale — operator HOLD active (sharp-bee consolidated lane); not pushing from silent-lynx-25 until cleared.

— sent from silent-lynx-25

[gunbai-bot]

Superseded by consolidate/int-literal-body-producer (clean base 12bc08a, drops the f956b1c/241a5cea2 regression). Branch preserved; parse-side work + pick witnesses reused in the consolidated lane.