Slate v2 Single Roadmap Consolidation

Supporting plan. For current queue and roadmap truth, see master-roadmap.md.

Goal

Make the doc stack stop fighting itself:

• cohesive-program-plan.md owns numbered program phases
• package-end-state-roadmap.md owns package-level sequencing and next slices
• downstream docs and slate-v2 repo docs should point at those owners instead of restating competing roadmap language
• patch the stale endgame phase model so the docs stop pretending all remaining work is one giant Phase 7 blob

Plan

1. Audit plate-2 and .tmp/slate-v2 for duplicated roadmap and status language.
2. Define the exact ownership split for phases, package sequencing, migration, and compatibility docs.
3. Edit docs to remove duplicated roadmap claims and point back to the owning doc.
4. Verify formatting and record remaining follow-up drift.

Findings

• cohesive-program-plan.md clearly owns Phases 0-7.
• package-end-state-roadmap.md clearly owns package-level done/next/later, but it needed an explicit ownership section so people stop treating it like a vague appendix.
• overview.md and final-synthesis.md both restate concrete next execution order in enough detail to drift from the package roadmap.
• phase7-migration-story.md and phase7-compatibility-envelope.md are fine as supporting docs, but they should not read like primary roadmap owners.
• .tmp/slate-v2/Readme.md still contains sequence language like "the next v2-native inline-family slice after mentions", which will rot.
• public .tmp/slate-v2 docs are allowed to describe the current proved surface, but they should not narrate future queue order.

Changes Made

• added explicit roadmap ownership rules to docs/slate-v2/package-end-state-roadmap.md
• renamed its queue section from Immediate Next Package Work to Current Package Queue
• updated docs/slate-v2/overview.md to summarize state and defer queue ownership to package-end-state-roadmap.md
• replaced the stale The Next Execution Order section in docs/slate-v2/final-synthesis.md with an execution hand-off rule
• updated docs/slate-v2/cohesive-program-plan.md so it owns phases, not the concrete package queue
• removed ordered inline-family queue language from docs/slate-v2/phase7-migration-story.md
• removed stale "first/next slice" wording from .tmp/slate-v2/Readme.md
• marked .tmp/slate-v2/docs/general/replacement-candidate.md and .tmp/slate-v2/site/examples/Readme.md as current-surface docs, not roadmap owners
• patched the stale endgame phase ladder in docs/slate-v2/cohesive-program-plan.md:
  • Phase 7 = historical public-surface cashout
  • Phase 8 = active package API shaping
  • Phase 9 = replacement-envelope expansion
  • Phase 10 = release-readiness gate
• updated downstream docs so they stop talking like Phase 7 is still the live bucket:
  • docs/slate-v2/engine.md
  • docs/slate-v2/overview.md
  • docs/slate-v2/final-synthesis.md
  • docs/slate-v2/phase6-hardening.md
  • docs/slate-v2/package-end-state-roadmap.md

Current Follow-On

• patch numbered phases so Phase 7 becomes historical cashout instead of the forever-current bucket
• update downstream docs that still mention "the next move is Phase 7"

Remaining Follow-Up

• phase7-compatibility-envelope.md and .tmp/slate-v2/docs/general/replacement-candidate.md both list the current envelope. That duplication is acceptable because one is internal and one is public, but they still need sync discipline.
• anchor-surface bullets are repeated across Readme.md, package READMEs, and examples docs. That is probably fine reference duplication, not urgent DRY debt.
• the phase7-* filenames are now historical labels. That is acceptable for now, but if the doc stack gets noisier later, rename or add a tiny historical note at the top instead of pretending those files are the active phase queue.

Progress

• Loaded planning/research skills.
• Audited top-level plate-2 docs and key .tmp/slate-v2 public docs for roadmap ownership drift.
• Edited the doc stack so phases live in cohesive-program-plan.md and package queue ownership lives in package-end-state-roadmap.md.
• Re-cut the endgame so the active phase is Phase 8 instead of the old fake forever-Phase-7 story.
• Verified formatting with Prettier in both repos.
• Verified the stale "next/first slice" phrases are gone from the audited docs.