[d6g8k5htny-coder]

Branched from tip 388a22caff13c449f0e3264b0aff314d456d6659 (after #97).

Scientific effect: NONE. OBL-H5-JETMOD stays OPEN. No flags flipped. SoT stays ABSENT. Engineering is not discharge.

Why a separate errata page

docs/OPEN_PROBLEMS.md is untouched because it is archive-pinned. The first pin is SUPPLEMENTAL_DEPENDENCIES in tools/twelve_project_check.py (28,417 B, sha256 8f404f87…dfb4d). The second is lambda_uniform/DEPENDENCIES.json inside the hash-pinned research/campaigns/q0_twelve_20260920_v1.zip. An in-place edit would break the CI replay, so the corrections live on a new page.

Changes

• New: docs/OPEN_PROBLEMS_FROZEN_ERRATA_20260925.md
  • E1 (A5 lines 116–120, SIDE24_CELL): RN_SIDE24_CELL.md claims only C=[1−1/4000,1+1/4000]^2, with an upper below 0.00000559 conditional on the imported H3 floor. The rectangle [1999/2000,2001/2000]^2, its four-cell split and the ≤3/500000000000 upper are the engineering replay of tools/rn_side24_spatial_check.py, not a claim of the note. Complement: the rectangle minus C, and the near annulus 0.1≤|y|≤5.
  • E2 (A5 lines 89–92): research/cover/ is a region-generic engineering engine, not the CL-RNU-001 Piece-2 driver. That driver stays REQUIRED CARRIER ABSENT, consistent with STATUS.md, STATUS_RN_UNIF.md and PACKET.json. The page adds an ABSENT/REPLACED crosswalk covering historical rnu_* ABSENT, the rnu_white.py reconstruction as a separate node, orders 2–4 REPLACED on 5≤|y|≤17 at r=1/20 with its complement, successors S1 then S2, and "CERTIFIED q=2" as carrier absent (see docs(math_status): STATUS.md "CERTIFIED" q=2 envelope is REQUIRED CARRIER ABSENT (walk-down 9, M1) #101).
  • Links docs/DOWNSTREAM_RN_CROSSWALK_20260925.md (from docs(inventable): RN downstream crosswalk + mesoscopic challenge + Math- #8 handoff (fresh on 077464ef; supersedes #92) #97).
• Edited: docs/NAVIGATION.json, which gains one page entry.

Not touched: OPEN_PROBLEMS.md, STATUS*, PACKET.json, flags and observe fields. Agent 6 will add a pointer later.

Local checks (on 388a22c + this diff)

These all passed: navigation_check (15 pages, 0 problems), math_status_check, consumers_check, mirror_quotes_check, claims_check, lanes_check, twelve_project_check (py3.11: PASS 12/12), and pytest (navigation, math_status, claims, consumers, lanes, mirror_quote*, mirrors_index: 266 passed / 2 skipped; test_twelve_project_check: 54 passed).

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.