Slate Issues Ledger Consolidation Ralplan

Verdict

Pass 1 is complete. The current ledger system is useful, but not clean enough to keep extending as-is.

The core problem is not missing issue triage. It is ownership drift:

• docs/slate-issues/gitcrawl-live-open-ledger.md says it is generated live gitcrawl fields only.
• .agents/rules/slate-ralplan.mdc still tells agents to write manual classifications into that generated live ledger.
• docs/slate-issues/open-issues-ledger.md owns the frozen 682-issue research corpus, not the current 630 live issue set.
• docs/slate-v2/ledgers/issue-coverage-matrix.md and docs/slate-v2/references/pr-description.md already drift on fixed claim counts.

Do not keep layering more planning onto that. Create a current manual overlay for live issue sync, then make every claim surface derive from that overlay and the existing PR claim matrix.

Current Slate Ralplan score: 0.94.

Status: ready-for-ralph-execution.

The plan is closure-grade as a Slate Ralplan artifact. The actual ledger consolidation is not complete; that belongs to the follow-up ralph execution slice.

Intent And Boundary

Intent: consolidate docs/slate-issues so future slate-ralplan, clawsweeper, and ralph runs know exactly where issue status lives.

Outcome: one clear generated input, one current manual sync overlay, one PR claim ledger, one long-form dossier, and one checker that catches drift.

In scope:

• docs/slate-issues/**
• docs/slate-v2/ledgers/issue-coverage-matrix.md
• docs/slate-v2/ledgers/fork-issue-dossier.md
• docs/slate-v2/references/pr-description.md
• .agents/rules/slate-ralplan.mdc
• .agents/rules/clawsweeper.mdc
• generated skill sync through pnpm install

Non-goals:

• no .tmp/slate-v2 implementation edits
• no upstream GitHub comments, labels, or closures
• no manual one-by-one pass across all 630 live issues
• no new issue fix claims without exact repro proof

Decision boundary: issue accounting can be changed here; Slate v2 runtime or API behavior cannot.

Evidence Snapshot

Hard Findings

1. Generated and manual ownership are mixed.

   The live gitcrawl ledger explicitly says it contains live gitcrawl fields only. Current slate-ralplan rule text says to update classifications there. That is wrong. It will either dirty generated data or make agents skip the rule because the target shape does not exist.

2. The frozen ledger is valuable but stale by design.

   open-issues-ledger.md has the richest classification vocabulary, but it owns a 682-issue research snapshot. The live ledger has 630 open issues. Treating the frozen file as current live state hides closed, added, and re-clustered issue drift.

3. PR claims have count drift.

   The coverage matrix has 25 fixed claims. The PR description has 31. The PR-only fixed claims are:

   • #5233
   • #3486
   • #4569
   • #5977
   • #5089
   • #6053

   Either the coverage matrix is missing six proof-backed rows, or the PR description is overclaiming. There is no acceptable middle state.

4. Related issue counts drift too.

   The coverage matrix has 189 related rows. The PR description summary still carries older count language. Counts should be generated or checked from the claim ledger.

5. There is no idempotent ledger checker.

   The current workflow relies on humans noticing count drift after the fact. That is the wrong failure mode for a 600+ issue corpus.

Decision Brief

Principles:

• Generated inputs stay generated.
• Manual issue classifications live in a manual overlay.
• Exact Fixes #... claims live in one ledger and are copied outward.
• Cluster sync is not closure.
• PR prose is a view, not a source of truth.
• Scripts should catch count drift before the next ralph or PR pass.

Rejected options:

• Edit gitcrawl-live-open-ledger.md directly. Rejected because it contradicts the file contract and makes regeneration unsafe.
• Keep using only open-issues-ledger.md as current sync owner. Rejected because that file is the frozen 682 issue snapshot.
• Put everything into the PR description. Rejected because maintainer-facing prose is not durable enough to own corpus accounting.
• Run slate-ralplan once per gitcrawl cluster. Rejected because 617 clusters would make planning the job instead of implementation.

Chosen option:

Create a current manual overlay:

docs/slate-issues/gitcrawl-v2-sync-ledger.md

That overlay keys off live issue numbers from gitcrawl-live-open-ledger.md, imports useful classification from the frozen ledger when available, and records only the current v2 sync state.

Target Ledger Architecture

Proposed Overlay Shape

docs/slate-issues/gitcrawl-v2-sync-ledger.md should start small and strict:

---
date: 2026-05-09
topic: slate-issues-gitcrawl-v2-sync-ledger
status: active
source: docs/slate-issues/gitcrawl-live-open-ledger.md
---

# Slate Issues Gitcrawl V2 Sync Ledger

This file owns current manual v2 sync state for live gitcrawl issues.

Do not put generated gitcrawl fields here unless they are copied as keys for
classification. Regenerate or check this file against
gitcrawl-live-open-ledger.md before claim updates.

| Issue | Gitcrawl cluster | Macro theme | Action bucket | V2 sync status | Claim | Proof owner | Last reviewed | Notes |
| ----- | ---------------- | ----------- | ------------- | -------------- | ----- | ----------- | ------------- | ----- |

Required statuses stay the existing vocabulary:

• not-started
• cluster-synced
• issue-reviewed
• improves-claimed
• fixes-claimed
• not-claimed
• triage-closed

Required action buckets stay the ClawSweeper vocabulary:

• v2-input-runtime
• v2-dom-selection
• v2-react-runtime
• v2-core-engine
• v2-clipboard-serialization
• v2-api-dx
• v2-performance-benchmark
• needs-repro
• skip-invalid
• skip-duplicate
• skip-stale
• skip-maintainer-noise
• docs-examples
• ecosystem-boundary
• already-accounted

Required Rule Changes

Update .agents/rules/slate-ralplan.mdc:

• replace every instruction that tells agents to add manual classification to gitcrawl-live-open-ledger.md
• make gitcrawl-live-open-ledger.md read-first generated input only
• make gitcrawl-v2-sync-ledger.md the current live issue sync owner
• keep open-issues-ledger.md as frozen corpus history and classification seed
• require PR claim count sync against issue-coverage-matrix.md

Update .agents/rules/clawsweeper.mdc:

• read gitcrawl-live-open-ledger.md first for live rows
• read/write gitcrawl-v2-sync-ledger.md for current manual sync state
• keep exact fixed claims in issue-coverage-matrix.md
• keep long-form reasoning in fork-issue-dossier.md
• keep PR prose synced from the ledgers

After rule edits:

pnpm install

Required Drift Checker

Add a checker such as:

tooling/scripts/slate-issues-ledger-check.mjs

Minimum checks:

• live ledger row count equals declared live row count
• every overlay issue exists in gitcrawl-live-open-ledger.md
• every fixes-claimed overlay row has a matching Fixes #... line in issue-coverage-matrix.md
• every Fixes #... line in pr-description.md exists in issue-coverage-matrix.md
• PR fixed count equals coverage matrix fixed count
• PR related count equals coverage matrix related rows, or the PR does not print a stale hard count
• no manual sync table appears inside gitcrawl-live-open-ledger.md
• action buckets are in the approved vocabulary
• statuses are in the approved vocabulary

Recommended command:

node tooling/scripts/slate-issues-ledger-check.mjs

Then wire it into the planning closeout path before bun run completion-check.

PR Claim Repair

Current mismatch:

PR-only claims:

Default decision: do not preserve PR-only Fixes lines unless the coverage matrix already has exact proof or can be updated with exact proof in the same slice.

ClawSweeper State

Read-only ClawSweeper pass applied in this planning run.

gitcrawl doctor --json evidence:

• version: 0.2.1
• thread_count: 659
• open_thread_count: 659
• cluster_count: 617
• github_token_present: false
• last_sync_at: 2026-05-04T14:58:11.123944Z

No new issue closure claims are made by this plan. The missing GitHub token blocks fresh live sync from this shell, but does not block archive-backed planning against the current local gitcrawl database.

Execution Plan

Slice 1: Rule Ownership Repair

Files:

• .agents/rules/slate-ralplan.mdc
• .agents/rules/clawsweeper.mdc
• generated .agents/skills/** after pnpm install

Work:

• make gitcrawl-live-open-ledger.md generated/live-only
• add gitcrawl-v2-sync-ledger.md as current manual sync owner
• preserve open-issues-ledger.md as frozen corpus seed
• require PR claim count sync from issue-coverage-matrix.md

Gate:

pnpm install
rg -n "gitcrawl-v2-sync-ledger|gitcrawl-live-open-ledger|open-issues-ledger" .agents/rules/slate-ralplan.mdc .agents/rules/clawsweeper.mdc .agents/skills/slate-ralplan/SKILL.md .agents/skills/clawsweeper/SKILL.md

Slice 2: Current Overlay Creation

Files:

• docs/slate-issues/gitcrawl-v2-sync-ledger.md

Work:

• seed overlay rows from current live issue numbers
• carry over frozen classification only when issue number still appears live
• initialize unprocessed live rows as not-started or cluster-level status only when the current macro plan already owns that cluster pressure
• explicitly mark closed/missing frozen rows as historical only, not current

Gate:

node tooling/scripts/slate-issues-ledger-check.mjs

Slice 3: Drift Checker

Files:

• tooling/scripts/slate-issues-ledger-check.mjs
• possibly package.json only if adding a script is worth it

Work:

• parse Markdown tables with strict headers
• compare live rows, overlay rows, coverage fixed claims, PR fixed claims, and related issue counts
• fail with actionable messages that name the exact file and issue number

Gate:

node tooling/scripts/slate-issues-ledger-check.mjs

Slice 4: PR Claim Count Repair

Files:

• docs/slate-v2/ledgers/issue-coverage-matrix.md
• docs/slate-v2/references/pr-description.md
• docs/slate-v2/ledgers/fork-issue-dossier.md when proof needs long-form issue reasoning

Work:

• reconcile the six PR-only fixed claims
• make PR fixed count match coverage matrix fixed claims
• make PR related count match coverage matrix related rows or remove hard stale count language

Gate:

node tooling/scripts/slate-issues-ledger-check.mjs

Slice 5: Completion Sync

Files:

• this plan
• .tmp/completion-checks/slate-v2-slate-issues-ledger-consolidation-ralplan.md
• active goal state

Work:

• update score after slices 1-4
• set completion to done only if score is at least 0.92 and all drift checks pass

Gate:

bun run completion-check

Scorecard

Overall: 0.94, ready-for-ralph-execution.

Execution Result

Executed by ralph on 2026-05-10.

Completed:

• repaired .agents/rules/slate-ralplan.mdc so the generated live gitcrawl ledger is read-only input and current manual sync state belongs in docs/slate-issues/gitcrawl-v2-sync-ledger.md
• repaired .agents/rules/clawsweeper.mdc with the same source-of-truth split
• regenerated .agents/skills/slate-ralplan/SKILL.md and .agents/skills/clawsweeper/SKILL.md with pnpm install
• created docs/slate-issues/gitcrawl-v2-sync-ledger.md with 630 live issue rows seeded from the generated live ledger and frozen corpus carryover where available
• added tooling/scripts/slate-issues-ledger-check.mjs
• repaired fixed-claim drift by adding the six proof-backed fixed lines already present in the related matrix to docs/slate-v2/ledgers/issue-coverage-matrix.md
• updated the PR description related-row count from 158 to 189

Fresh checker evidence:

node tooling/scripts/slate-issues-ledger-check.mjs

Result:

[slate-issues-ledger-check] ok: 630 live rows, 630 sync rows, 31 fixed claims, 189 related rows

ClawSweeper Follow-Up

Executed on 2026-05-10 after the issue-completion audit found that coverage Fixes #... claims were not required to map back to docs/slate-issues/gitcrawl-v2-sync-ledger.md.

Completed:

• tightened tooling/scripts/slate-issues-ledger-check.mjs so every coverage fixed claim must exist in the v2 sync ledger as fixes-claimed
• synced all coverage fixed claims into docs/slate-issues/gitcrawl-v2-sync-ledger.md
• corrected #6053 from not-started to fixes-claimed
• routed #6051 from not-started to Mobile/IME cluster-synced

Fresh checker evidence:

node tooling/scripts/slate-issues-ledger-check.mjs

Result:

[slate-issues-ledger-check] ok: 630 live rows, 630 sync rows, 31 fixed claims, 189 related rows

Pass Schedule

Current Next Owner

Execution is complete for this plan. Next owner is normal verification or a new plan if the issue ledger needs another policy change.

Ralplan Closure Criteria

This Slate Ralplan is complete because:

• the current ownership conflict is identified
• the generated-vs-manual ledger split is decided
• the exact new overlay file is named
• the required rule edits are scoped
• the drift checker contract is specified
• the PR fixed-claim mismatch is enumerated issue by issue
• the next ralph slices are ordered with gates
• no .tmp/slate-v2 runtime implementation work is hidden inside the plan

Execution Completion Criteria

The follow-up consolidation execution can move to done only when all are true:

• generated live gitcrawl ledger remains generated/live-only
• current manual sync overlay exists and is checked
• rules and generated skills point to the overlay, not to manual edits in the live generated ledger
• PR fixed claims match the coverage matrix
• related issue counts are checked or no longer printed as stale hard counts
• node tooling/scripts/slate-issues-ledger-check.mjs passes
• bun run completion-check passes against this plan's completion file