Skip to content

FE-1089: Sync planning docs at orchestrator-cutover close#284

Open
kostandinang wants to merge 1 commit into
graphite-base/284from
ka/fe-1089-planning-sync
Open

FE-1089: Sync planning docs at orchestrator-cutover close#284
kostandinang wants to merge 1 commit into
graphite-base/284from
ka/fe-1089-planning-sync

Conversation

@kostandinang

@kostandinang kostandinang commented Jul 2, 2026

Copy link
Copy Markdown
Contributor

Stack Context

Final PR of the executor cutover stack (#274#282). It updates only the planning documents to match what the stack actually built. No product behavior changes — #282 below it stays a pure feature diff.

What?

  • memory/SPEC.md — the spec still described real execution as "inactive until the real-execution stack lands". That stack has now landed, so the executor decision (D101-L) and invariant (I56-L) are rewritten to describe reality: all execute_* tools are active in CODE mode, real git/subprocess/agent work enters only through the injected ExecutionPorts, and host changes are file-only behind an accepted commit SHA. Five per-PR implementation notes were collapsed into a pointer to src/executor/TOPOLOGY.md, which owns that detail. Also adds an ExecutionPorts glossary entry and names a known gap: the agent-runner tests prove wiring, not real-model output quality.
  • memory/PLAN.md — marks the orchestrator-cutover arc done, moves the six finished work-item definitions to docs/archive/PLAN_HISTORY.md, and records that FE-1116 (FE-1116: Session-local elicitation gaps from a graph-derived seed #280) and FE-1115 (FE-1115: Refine Pi TUI component DX: preview harness, component refinements, and new components #281) are in flight on parallel branches, so their own PLAN/SPEC edits stay authoritative when merges meet.
  • memory/cards/ — deletes 11 scope cards whose work is finished.
  • Two small ride-alongs, both doc-shaped: refreshed ownership text in src/executor/TOPOLOGY.md, and a code comment in active-tools.ts that cited the wrong invariant id (I52-L → I56-L).

Why?

The planning docs lagged the code: they described the executor as mostly inactive and listed six finished work items as live. Syncing them now gives the next piece of work (executor observability) accurate docs to plan against, and keeping the cleanup out of #282 keeps that PR reviewable on its own.

Verification

npm run verify — 1156 tests passed / 1 skipped; build and markdown-link checks green.

@cursor

cursor Bot commented Jul 2, 2026

Copy link
Copy Markdown

PR Summary

Low Risk
Documentation and comment-only edits; no runtime, auth, or execution logic changes in the shown diff.

Overview
Closes out orchestrator-cutover planning debt: 11 finished scope cards under memory/cards/ (agent-runner, host-promotion, promotion, plan-check) are removed so active planning no longer points at completed slices.

src/executor/TOPOLOGY.md is brought in line with landed behavior: SPEC pointers now cite D101-L / I56-L, the owns blurb states real git/subprocess/agent work enters only via injected ExecutionPorts, and the side-effect paragraph documents I56-L (one declared side effect per helper, port failure does not advance metadata, host mutation limited to accepted-SHA apply).

active-tools.ts comment only: execute-mode tool range FE-1089..FE-1118 and invariant reference I52-L → I56-L (no allowlist change).

Per PR intent, companion updates to memory/SPEC.md / memory/PLAN.md (executor active, arc done, archive pointers) are part of the same doc sync; they are not shown in the diff excerpt here.

Reviewed by Cursor Bugbot for commit dc7bfd0. Bugbot is set up for automated code reviews on this repo. Configure here.

kostandinang commented Jul 2, 2026

Copy link
Copy Markdown
Contributor Author

Warning

This pull request is not mergeable via GitHub because a downstack PR is open. Once all requirements are satisfied, merge this PR as a stack on Graphite.
Learn more

This stack of pull requests is managed by Graphite. Learn more about stacking.

@kostandinang kostandinang changed the base branch from ka/fe-1118-executor-host-promotion to graphite-base/284 July 3, 2026 16:52
@kostandinang kostandinang force-pushed the ka/fe-1089-planning-sync branch from a83bbae to 72e1db4 Compare July 3, 2026 16:53
@kostandinang kostandinang changed the base branch from graphite-base/284 to ka/fe-1118-executor-host-promotion July 3, 2026 16:53
ln-sync at arc close: record the completed D101-L port cutover in SPEC
(D101-L row, I56-L rewrite, requirement 26, ExecutionPorts lexicon row,
worker content-quality blind spot), thin the five D101-L refinement
blockquotes to src/executor/TOPOLOGY.md pointers, archive done frontier
definitions to PLAN_HISTORY, close the orchestrator-cutover arc in PLAN,
acknowledge in-flight FE-1116/FE-1115 (PRs #280/#281), reconcile executor
topology headers, fix the stale I52-L comment to I56-L, and delete 11
exhausted scope cards.
@kostandinang kostandinang changed the base branch from ka/fe-1118-executor-host-promotion to graphite-base/284 July 3, 2026 16:59
@kostandinang kostandinang force-pushed the ka/fe-1089-planning-sync branch from 72e1db4 to dc7bfd0 Compare July 3, 2026 17:00
@kostandinang kostandinang changed the base branch from graphite-base/284 to ka/fe-1118-executor-host-promotion July 3, 2026 17:00
@kostandinang kostandinang changed the base branch from ka/fe-1118-executor-host-promotion to graphite-base/284 July 3, 2026 17:41
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants