v2 complexity.dag rework — well-modeled lens applying v4 modeling discipline (CALIBRATION investigation, brief at /tmp/v2_complexity_lens_brief.md)

[briansrls]

Auto-opened by session-dashboard for session deep-seal-431.
Pushing to session/deep-seal-431 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.

[briansrls]

Review metadata

• Provider / model: codex / unknown
• Commit: 1f8d55e2 · Trigger: schedule
• Thinking: 269s wall

---

BLOCKING (3)

Root Cause

• src/v2/complexity.dag asymptotic class is being modeled as a new local enum rather than projected through std.induction.CostBound/ProductBound → reuse the existing bound carrier or add a valid dissolution-tagged bridge with a concrete trigger.
• src/v2/complexity.dag ComplexityBound lacks structural product/log-factor coordinates even though CostBound already has ProductBound factors → represent polynomial and log factors structurally or keep classification on normalized CostExpr/CostBound.
• src/v2/complexity.dag near-miss semantics are encoded both in discharge_complexity_lens_tdd_case and again at the table call site → keep the expectation semantics in one place and call the case directly.

⚠️ The PR introduces a duplicated, under-modeled complexity-bound carrier and a concrete false TDD table result, so it needs rework before merge.

[briansrls]

BLOCKING: ComplexityBound adds a new substrate coproduct without the required 🟢/🟡/🔴 tag and duplicates the existing std.induction.CostBound asymptotic-bound algebra instead of grounding in it (INVARIANTS P1/P5, MODELING M9).

[briansrls]

BLOCKING: The new multiplication projection underestimates O(n * log n): BoundLinear times BoundLog falls through to BoundLog instead of BoundLinearithmic, so classify_complexity can report a lower class than CostExpr proves (INVARIANTS P1/P3).

[briansrls]

BLOCKING: complexity_lens_tdd_table_passes negates the near-miss case even though ExpectNearMiss already discharges as !matches, making the table fail when the near-miss behaves correctly.

[briansrls]

Closing per operator wrap-up directive 2026-05-20. Worker deep-seal-431 is archived; closing on their behalf as PM. Branch stays in place.