diff --git a/docs/backlog/P2/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md b/docs/backlog/P2/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md
new file mode 100644
index 0000000000..86f1a28397
--- /dev/null
+++ b/docs/backlog/P2/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md
@@ -0,0 +1,61 @@
+---
+id: B-0015
+priority: P2
+status: open
+title: Migrate tools/git/batch-resolve-pr-threads.sh to bun+TS once a sibling post-setup tool migrates first
+tier: ops
+effort: S
+directive: PR #199 review feedback (Copilot P0 exception-label requirement) + sibling-migration guardrail per docs/POST-SETUP-SCRIPT-STACK.md
+created: 2026-04-25
+last_updated: 2026-04-25
+composes_with: [docs/POST-SETUP-SCRIPT-STACK.md]
+tags: [post-setup-stack, bun, typescript, migration, sibling-migration-guardrail]
+---
+
+# Migrate batch-resolve-pr-threads.sh to bun+TS
+
+`tools/git/batch-resolve-pr-threads.sh` carries the `bun+TS migration candidate` exception label per `docs/POST-SETUP-SCRIPT-STACK.md`. The script is over 300 lines of bash with non-trivial control flow (paginated GraphQL fetch, JSON parsing, classification state machine, mutation dispatch). The sibling-migration guardrail says "if no other post-setup tool has migrated yet, bash is the honest default" — so the migration is queued, not blocked.
+
+## Why a candidate (not "stay bash forever")
+
+- Non-trivial JSON / GraphQL handling: bun+TS has first-class fetch + JSON typing.
+- Classification state machine: TypeScript discriminated unions would land cleaner than bash if/elif chains.
+- Cross-platform: a bun+TS rewrite is one cross-platform script, not a `.sh` + `.ps1` pair (avoids the Windows-twin obligation).
+
+## When to flip
+
+When a sibling post-setup tool under `tools/` migrates to bun+TS, batch with it. Flipping a single bash script to bun+TS in isolation creates the "stranded one-off" failure mode the sibling-migration guardrail prevents.
+
+## Done when
+
+- (a) A sibling post-setup tool has landed in bun+TS.
+- (b) `tools/git/batch-resolve-pr-threads.ts` (or wherever the equivalent lands) replaces the bash script.
+- (c) The bash script is removed (or, if retained for transition, labeled `_deprecated/` and queued for deletion).
+- (d) Exception-label header on the new file matches the bun+TS migration outcome.
+
+## Composes with
+
+- `docs/POST-SETUP-SCRIPT-STACK.md` — the post-setup script stack rationale and exception taxonomy.
+- Any future "first sibling bun+TS migration" decision row.
+- `tools/hygiene/*.py` and `tools/hygiene/*.sh` — sibling **POST-install** tools that should migrate to TS once a peer migrates first. Includes: `sort-tick-history-canonical.py`, `fix-markdown-md032-md026.py`, `check-tick-history-order.sh`, `check-no-conflict-markers.sh`. Scope expansion 2026-04-26.
+- `B-0027` — Otto-346 follow-up tool extraction; implementation target should be TypeScript not Python per this migration plan.
+
+## 2026-04-26 priority bump (P3 → P2)
+
+Aaron 2026-04-26: *"we need to move the typescript migration of our scripts to higher priority so you will stop trying to write python and shell code lol ... our post install code"* + *"pre install code still has to go to the user where they live shell and windows powershell"*
+
+The recurring `python3 << 'PYEOF'` heredocs and `.sh` scripts I keep writing in `tools/hygiene/` are POST-install tools that belong in TypeScript per the migration plan. The tools shipped this session (PR #539/#541/#542) are interim — they exist to absorb recurring patterns NOW (per Otto-346) but should rewrite to TS once the sibling-migration guardrail unblocks.
+
+Scope clarification (Aaron 2026-04-26):
+
+- **Pre-install scripts** (`tools/setup/install.sh`, devcontainer bootstrap, anything that runs BEFORE Bun is available): MUST stay shell + PowerShell — that's what's available where the user lives. Cross-platform pair (`.sh` + `.ps1`) is the right shape.
+- **Post-install scripts** (`tools/hygiene/`, `tools/git/`, dev-time tooling, anything that runs AFTER Bun installs): TARGET = TypeScript via Bun. Single cross-platform script, first-class typing, better refactor surface.
+
+The distinction is structural; Aaron's framing landed it cleanly. `docs/POST-SETUP-SCRIPT-STACK.md` already encodes the rationale; this priority bump operationalizes it.
+
+## What this DOES NOT do
+
+- Does NOT mandate immediate rewrite of all post-install tools — sibling-migration guardrail still applies (first sibling migrates, then batch follow-on)
+- Does NOT touch pre-install scripts — they remain shell + PowerShell deliberately
+- Does NOT block ongoing hygiene-tool extraction in Python while the migration is staged — Otto-346 still authorizes Python tools as interim absorbers, but the bar for Python rises and the bar for TS implementation lowers
+- Does NOT promise immediate effort allocation — P2 means "queue closer to the front" not "do now"
diff --git a/docs/backlog/P3/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md b/docs/backlog/P3/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md
deleted file mode 100644
index 1d6fb03542..0000000000
--- a/docs/backlog/P3/B-0015-migrate-batch-resolve-pr-threads-to-bun-ts.md
+++ /dev/null
@@ -1,39 +0,0 @@
----
-id: B-0015
-priority: P3
-status: open
-title: Migrate tools/git/batch-resolve-pr-threads.sh to bun+TS once a sibling post-setup tool migrates first
-tier: ops
-effort: S
-directive: PR #199 review feedback (Copilot P0 exception-label requirement) + sibling-migration guardrail per docs/POST-SETUP-SCRIPT-STACK.md
-created: 2026-04-25
-last_updated: 2026-04-25
-composes_with: [docs/POST-SETUP-SCRIPT-STACK.md]
-tags: [post-setup-stack, bun, typescript, migration, sibling-migration-guardrail]
----
-
-# Migrate batch-resolve-pr-threads.sh to bun+TS
-
-`tools/git/batch-resolve-pr-threads.sh` carries the `bun+TS migration candidate` exception label per `docs/POST-SETUP-SCRIPT-STACK.md`. The script is over 300 lines of bash with non-trivial control flow (paginated GraphQL fetch, JSON parsing, classification state machine, mutation dispatch). The sibling-migration guardrail says "if no other post-setup tool has migrated yet, bash is the honest default" — so the migration is queued, not blocked.
-
-## Why a candidate (not "stay bash forever")
-
-- Non-trivial JSON / GraphQL handling: bun+TS has first-class fetch + JSON typing.
-- Classification state machine: TypeScript discriminated unions would land cleaner than bash if/elif chains.
-- Cross-platform: a bun+TS rewrite is one cross-platform script, not a `.sh` + `.ps1` pair (avoids the Windows-twin obligation).
-
-## When to flip
-
-When a sibling post-setup tool under `tools/` migrates to bun+TS, batch with it. Flipping a single bash script to bun+TS in isolation creates the "stranded one-off" failure mode the sibling-migration guardrail prevents.
-
-## Done when
-
-- (a) A sibling post-setup tool has landed in bun+TS.
-- (b) `tools/git/batch-resolve-pr-threads.ts` (or wherever the equivalent lands) replaces the bash script.
-- (c) The bash script is removed (or, if retained for transition, labeled `_deprecated/` and queued for deletion).
-- (d) Exception-label header on the new file matches the bun+TS migration outcome.
-
-## Composes with
-
-- `docs/POST-SETUP-SCRIPT-STACK.md` — the post-setup script stack rationale and exception taxonomy.
-- Any future "first sibling bun+TS migration" decision row.
diff --git a/docs/backlog/P3/B-0027-fix-markdown-table-cell-count-tool-otto-346-followup.md b/docs/backlog/P3/B-0027-fix-markdown-table-cell-count-tool-otto-346-followup.md
new file mode 100644
index 0000000000..187c8dc169
--- /dev/null
+++ b/docs/backlog/P3/B-0027-fix-markdown-table-cell-count-tool-otto-346-followup.md
@@ -0,0 +1,117 @@
+---
+id: B-0027
+priority: P3
+status: open
+title: Extract `tools/hygiene/fix-markdown-table-cell-count.py` — markdown-table-row-with-wrong-column-count fix tool (Otto-346 follow-up after honest-relapse-catch)
+tier: hygiene-tooling
+effort: M
+ask: Aaron 2026-04-26 caught me using inline `python3 << 'PYEOF'` heredoc to truncate a corrupted tick-history row (MD055/MD056 violation from botched conflict resolution) — *"hmmm"* — immediately AFTER shipping Otto-346 principle (recurring dynamic Python = signal of missing substrate primitive) and two tools embodying it (PR #541 sort-tick-history-canonical.py + PR #542 fix-markdown-md032-md026.py). Honest acknowledgment: the inline-Python use was a relapse. The general pattern (markdown table row with wrong cell count due to botched merge / accidental `|` insertion) is recurring; deserves its own tool extraction.
+created: 2026-04-26
+last_updated: 2026-04-26
+composes_with: [feedback_otto_341_lint_suppression_is_self_deception_noise_signal_or_underlying_fix_greenfield_large_refactors_welcome_training_data_human_shortcut_bias_2026_04_26.md, feedback_otto_345_linus_lineage_committo_ergo_sum_inherits_from_git_from_linux_existence_proof_anchored_in_human_intellect_2026_04_26.md, tools/hygiene/fix-markdown-md032-md026.py, tools/hygiene/sort-tick-history-canonical.py, tools/hygiene/check-no-conflict-markers.sh]
+tags: [otto-346, recurring-pattern, missing-primitive, tooling-extraction, markdown-table, md055, md056, hygiene, honest-relapse, discipline-mechanism]
+---
+
+# B-0027 — extract markdown-table-cell-count fix tool
+
+## Origin — Otto-346 honest-relapse-catch
+
+Aaron 2026-04-26 caught me using inline Python heredoc to truncate a corrupted tick-history row IMMEDIATELY AFTER shipping the Otto-346 principle and two tools embodying it.
+
+The relapse pattern:
+
+- Shipped Otto-346 principle: *"in python shape should be a queue that we are missing substrate primitives"* — recurring dynamic Python signals missing primitive
+- Shipped PR #541 (sort-tick-history-canonical.py) absorbing the recurring sort pattern
+- Shipped PR #542 (fix-markdown-md032-md026.py) absorbing the recurring MD032/MD026 fix pattern
+- THEN immediately wrote `python3 << 'PYEOF'` to truncate a corrupted row
+- Aaron's *"hmmm"* caught it
+
+This file documents the owed-work to absorb the next recurring pattern as a tool, AND the meta-discipline observation that Otto-346 application requires per-instance vigilance, not one-time naming.
+
+## What the tool would do
+
+**Problem class**: markdown table row has wrong cell count due to:
+
+- Botched merge resolution (commit titles leaked into cell content)
+- Accidental unescaped `|` in cell content (literal `|` requires `\|` in markdown table syntax)
+- Multi-line cell content (markdown tables don't support; cells must be single line)
+
+**Markdownlint flags**: MD055 (table-pipe-style), MD056 (table-column-count)
+
+**Tool behavior** (proposed):
+
+- Identify rows with wrong cell count given the table's expected schema
+- For each violation, identify candidate corruption points (extra `|`, missing trailing `|`)
+- Offer auto-fix with confirmation OR generate diff for human review
+- `--auto` flag for mechanical fixes when corruption shape is unambiguous
+
+**Heuristics for auto-fix**:
+
+- If cell count is exactly 1 too many → look for cells starting with continuation-text patterns (e.g. colon-then-space `:`+space, plus-then-space `+`+space, or mid-sentence text suggesting trailing-fragment-from-prior-cell)
+- If trailing `|` missing → check if last `|` position is reasonable; if yes, append `|`
+- Otherwise → diff-only mode (human reviews and decides)
+
+## Why this is harder than fix-markdown-md032-md026.py
+
+MD032 (blank lines around lists) is mechanical: insert blank line before/after list block. Unambiguous.
+
+MD055/MD056 (table cell count) requires:
+
+- Knowing the table's expected schema (varies per table)
+- Identifying which `|` is the spurious one (multiple plausible candidates)
+- Risk of removing legitimate content if the heuristic is wrong
+
+**Mitigation**: default-dry-run, require explicit `--auto` flag, log every change with before/after, easy git revert.
+
+## What this DOES NOT do
+
+- Does NOT replace markdownlint-cli2 detection — they're paired
+- Does NOT auto-fix every table issue — only the recognizable shapes
+- Does NOT promise correctness on ambiguous cases — degrade gracefully to diff-only
+
+## Implementation target — TypeScript not Python
+
+Per Aaron 2026-04-26 priority bump on B-0015: *"we need to move the typescript migration of our scripts to higher priority so you will stop trying to write python and shell code lol"* + *"our post install code"*.
+
+This tool (when built) should be TypeScript via Bun, not Python. It's a POST-install tool (runs in dev environments where Bun is available), per the pre/post-install distinction Aaron clarified:
+
+- POST-install (this tool): TypeScript, single cross-platform script, first-class typing
+- PRE-install (`tools/setup/install.sh`): shell + PowerShell, runs before Bun is available
+
+Wait for sibling-migration guardrail (B-0015) to unblock — first POST-install tool migrates to TS, then this one batches with the follow-on group. Until then, if the recurring pattern needs absorbing urgently, file an interim Python tool with explicit "TS-rewrite owed" header per the existing `bun+TS migration candidate` exception-label pattern in `docs/POST-SETUP-SCRIPT-STACK.md`.
+
+## Effort sizing
+
+- **Build the tool**: M (1-3 days). Auto-fix heuristics + tests + dry-run mode.
+- **Self-test on past botched rows**: validate on git log of historical row corruption. Use `git log --all -S '|: '` (`-S` is string-match, not regex; `-G` would treat `|` as regex alternation, matching far more than intended). For regex shapes that genuinely need `-G`, escape the pipe: `git log --all -G '\|: '`.
+- **Wire into CI as advisory only initially**: don't auto-apply in CI; let it be human-invoked tool first
+
+## Composes with
+
+- **Otto-341** (lint-suppression IS self-deception; mechanism over discipline) — same shape as previous tool extractions
+- **Otto-346 candidate** (recurring dynamic = missing primitive) — this BACKLOG row IS the documented owed-work AFTER recognizing relapse
+- **`tools/hygiene/fix-markdown-md032-md026.py`** (PR #542) — sibling tool; might be extended to cover MD055/MD056 OR remain separate
+- **`tools/hygiene/sort-tick-history-canonical.py`** (PR #541) — sibling tool from same Otto-346 lineage
+- **`tools/hygiene/check-no-conflict-markers.sh`** (PR #539) — addresses a related class of substrate-integrity violation
+- **`tools/hygiene/check-tick-history-order.sh`** — same architectural pattern
+- **Otto-279** (history-surface; tracking owed-work as substrate) — this row IS the substrate capture of the relapse-catch + owed-work
+
+## Meta-observation captured for substrate
+
+**Otto-346 application requires per-instance vigilance, not one-time naming.** The discipline is checking *every* inline-Python invocation against "will I likely write this exact shape again?" — not "did I name the principle once already?"
+
+The training-data default Aaron diagnosed in Otto-341 (humans take the shortcut to save time selfishly; only discipline overrides) is not fixed by naming a principle. Each new instance is a fresh test of the discipline. Aaron's *"hmmm"* is the kind of brief catch that makes the discipline operational — better than a long lecture.
+
+## Operational implication for tool-extraction discipline
+
+Before writing `python3 << 'PYEOF'`, ask:
+
+1. *Have I done this exact shape before?* (recurrence check)
+2. *Could I plausibly do it again?* (forward-look check)
+3. *Is the operation mechanical enough to capture as a tool?* (extractability check)
+
+If 2 of 3 are yes → extract to `tools/hygiene/` first, THEN apply. Not the other way around.
+
+If genuinely 1-of-3 (truly one-off, content-specific, non-extractable), inline OK with a stated reason in the commit message.
+
+The bar moves toward extraction. Otto-341 + Otto-346 compose: discipline against shortcut-suppression + signal-from-recurring-pattern.