Skip to content

docs(rfc): Ultra Mode task hierarchy redesign (on hold) - #34241

Closed
chenmingwei23 wants to merge 1 commit into
mainfrom
docs/rfc-ultra-mode-task-hierarchy
Closed

chenmingwei23 wants to merge 1 commit into
mainfrom
docs/rfc-ultra-mode-task-hierarchy

Conversation

@chenmingwei23

@chenmingwei23 chenmingwei23 commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

Problem / Motivation

Goal: Record the Ultra Mode task hierarchy redesign and its task plan in the repository, so the work can resume later from a written design.

Ultra Mode's work ledger is flat: one task maps to one session. A level of tasks exists only as another conductor ledger, a worker session holds one open item, nesting depth is capped at 2, and there is no way to say that one task waits on another or that several sessions share a task. Larger goals cannot be decomposed without folding hand-offs into a single item.

Why it matters

Without a written design, the three gaps (task tree, dependency edges, many-to-many assignment) have no agreed shape and the owner's rulings on the open questions are not recorded. The owner paused this project on 2026-11-09, so the document is the only place the design and the 7-PR plan (R0, F1-F4, U1, S1) will live until work resumes. Tracking issue: #34239.

Not a goal

  • Building any of the design. Nothing is built and no code changes.
  • Accepting the RFC. Its status stays draft and the project stays on hold.
  • Touching any other RFC row in the index beyond adding this one.

What changed (motivation -> approach -> change)

The RFC replaces the flat ledger with three structures: a task tree of any depth, dependency edges between tasks (cycle-checked, ready-set dispatch, critical path first), and many-to-many assignment of tasks to sessions. The principle is to write facts and derive everything: only facts go into the crew log, and the tree, roll-ups, ready set and caps are derived by the work fold at read time, so old ledgers read unchanged. Open questions Q1-Q6 are replaced by a rulings table.

  • Adds docs/request-for-change/rfc-ultra-mode-task-hierarchy.md (status draft, on hold).
  • Adds docs/request-for-change/plans/2026-11-08-ultra-mode-task-hierarchy.md and its row in docs/request-for-change/plans/README.md.
  • Adds the RFC's row and a one-sentence note to docs/request-for-change/README.md.
  • Rebased onto current main; the index conflict was resolved by keeping main's rows and adding this one.

Backwards compatibility

Compatible: documentation only, nothing that works today stops working.

Tests

Docs only. test/test_pr_quality_gates.py -k "rfc or request or index or front": 4 passed locally after the rebase.

Proof: not provable -- docs or config only

Manual verification

N/A -- docs only; the index row and links were checked by reading the rendered diff.

Related Issues

Tracking issue: #34239

Checklist

  • At most two commits (one is the norm), with a Conventional Commits title
  • Existing tests pass and new tests added for new functionality (N/A, docs only)
  • Self-review completed; code follows project style guidelines
  • Documentation updated (if applicable)
  • No secrets, credentials, or internal references in the diff

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

First Principles Review (Opus 5.5) — ⏭️ skipped

Revision 031039d5c6da97ae759c1bc6262621d3996860a3 ships no reviewable capability (docs, tests or generated files only), so there is nothing to inventory. Does not block merge.

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

UX Review (Opus 5.5) — ⏭️ skipped

Revision 031039d5c6da97ae759c1bc6262621d3996860a3 touches no user-facing surface (no changes under website/ or committed screenshots), so the UX review was skipped. Does not block merge.

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

Design Review (Opus 5.5) — ✅ PASS

Design-level review of 031039d5c6da97ae759c1bc6262621d3996860a3 — updated in place on each push. A BLOCK verdict blocks PR readiness; PASS does not.

Design-Verdict: PASS

Docs only: this adds a draft RFC and plan that are on hold, plus their index rows. No code, contract or user-facing surface changes.

[DESIGN-REVIEWED] 031039d

Description read

[DESCRIPTION-READ] 7039b6aba58f1534b8a9edd4e680c383c02f576a95ab71907f09c7bc066714ba

That sha256 names everything this verdict read: the PR description, plus the 1 evidence file(s) the workflow collected from it and handed this review. It covers that evidence because the media strip replaces every attachment URL with the same placeholder, so an edit swapping one attachment for another would leave the description's captured text identical while what this review judged changed. Recompute it with .github/scripts/pr-description-capture.sh over the same inputs; if it differs, the description or that evidence changed after this verdict was formed and any finding drawn from either is unproven.

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

GPT 6.1 Review — ✅ no blocking findings

GPT 6.1 completed its review of 031039d5c6da97ae759c1bc6262621d3996860a3 and found no blocking issues.

This comment is updated in place on each push.

Review details

No findings.
[GPT-REVIEWED] 031039d

False positive or not applicable? A repository writer can comment:
/ai-review override gpt 031039d5c6da97ae759c1bc6262621d3996860a3 false-positive: <reason, 20 to 500 characters>

@github-actions

github-actions Bot commented Nov 10, 2026 •

Copy link
Copy Markdown
Contributor

Opus 5.5 Review — ✅ no blocking findings

Reviewed 031039d5c6da97ae759c1bc6262621d3996860a3 — this comment is updated in place on each push.

Review details

No findings.

Notes:

  • The RFC's §1.2 problem 5 and its README row give DEFAULT_GOAL_CONCURRENCY_CAP = 6 and DEFAULT_GOAL_ITEM_CAP = 20. Main has DEFAULT_GOAL_CONCURRENCY_CAP = MAX_ITEMS_PER_CONDUCTOR (32) at work_ledger_records.py:147 and no item cap. Re-verify the values or mark them as true only at 8ad14ca919.
  • §11 says "the structural cap" bounds a hostile conductor, but the Q1 ruling ("No hard cap") removes that cap. §11 also names actions (add_child, dep_add) that are not in the §4 tool table.

[OPUS-REVIEWED] 031039d

Verdict parsed from the review's SHA-scoped output markers for commit 031039d5c6da97ae759c1bc6262621d3996860a3.

False positive or not applicable? A repository writer can comment:
/ai-review override fable 031039d5c6da97ae759c1bc6262621d3996860a3 false-positive: <reason, 20 to 500 characters>

@github-actions github-actions Bot added readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge and removed readiness: checking Automated validation is still running labels Nov 10, 2026
@chenmingwei23
chenmingwei23 force-pushed the docs/rfc-ultra-mode-task-hierarchy branch from a79f636 to 4f2588d Compare November 14, 2026 11:15
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: action required A blocking check or review needs attention merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 14, 2026
@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 14, 2026
@KiroCrewCaptain

Copy link
Copy Markdown
Collaborator

This adds a new concept, so it needs a design sign-off from a maintainer before it can merge. A maintainer will post the decision here.

To propose a design, follow the request-for-change process in CONTRIBUTING.md (docs/request-for-change/).

Review result: no blocking findings on this head. The change is docs only: a draft RFC, its task plan, and one index row in each of the two READMEs. PR Readiness is green on this commit. The code claims we checked hold at the stated audit commit 8ad14ca919, including WORK_MAX_DEPTH_DEFAULT = 2, DEFAULT_GOAL_CONCURRENCY_CAP = 6, WORK_CLOSED_WINDOW = 32 and worker_already_dispatched. Pinning every citation to one commit makes the RFC easy to verify, and that is well done.

Optional:

  • RFC section 7 (around line 408) and plan task F2 still say "item_cap counts leaves, with a higher default". Main dropped the goal item cap in refactor(work-ledger): drop the advisory goal item cap and its notice #33601 under the accepted rfc-drop-goal-item-cap.md. Consider dropping that clause and marking DEFAULT_GOAL_ITEM_CAP in section 1.2 row 5 as removed since the audit commit.
  • Section 11 (around line 453) says a structural cap bounds what a hostile conductor can build. Ruling Q1 says there is no hard cap, only a soft patrol signal. Consider naming the bounds that remain (the cycle check and the per-entry byte limit), or saying plainly that tree size has no bound beyond the soft signal. Also update section 3.1 to match Q1.
  • The RFC header says "Tier: T3", but tracking issue Ultra Mode task hierarchy: task tree, dependencies and many-to-many assignment (on hold) #34239 is labelled tier:T4. Consider changing the header to T4, or pointing to the issue label.

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 14, 2026
@chenmingwei23
chenmingwei23 force-pushed the docs/rfc-ultra-mode-task-hierarchy branch from 4f2588d to a1220e8 Compare November 14, 2026 22:26
@github-actions github-actions Bot added readiness: checking Automated validation is still running and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge labels Nov 14, 2026
@KiroCrewCaptain

KiroCrewCaptain commented Nov 14, 2026 •

Copy link
Copy Markdown
Collaborator

Code change approver recommendation: A person should decide

Why: the decider estimates a 12% chance that this change introduces a defect, within the 12% limit for approving on its own, but it changes 576 lines, more than the 400 it may approve on its own.

Docs-only PR that adds a draft RFC (on hold), its 7-PR plan and one index row in each of the two RFC indexes. It changes no code, config or tests, and all CI checks passed. §11 says the design adds no new trust boundary, which holds: every new action is a conductor-side work_ledger_record write behind the existing caller checks. Lifetime item creates are already unbounded on main. The remaining problems are stale or inconsistent wording, with no behaviour impact.

An automated recommendation, not a review: a maintainer still decides. It approves on its own only when the estimated chance of a defect is at most 12% and the change has at most 400 changed lines. Decider v3c (217fe607) · commit 031039d5c6 · 576 changed lines. If this recommendation is wrong, reply here: replies are how the approver is evaluated.

@github-actions github-actions Bot added readiness: passed Eligible automated validation passed for the current revision and removed readiness: checking Automated validation is still running labels Nov 14, 2026
@bolichen97

Copy link
Copy Markdown
Contributor

I read the whole diff at a1220e8. It's docs only: the RFC, its task plan, and an index row in each README.

The RFC marks itself on hold. The front matter says status: draft, the first line says "draft, on hold (paused by the owner on 2026-11-09)", and both README rows say "On hold" with nothing built. So I'm leaving a comment only, not a sign-off. The direction is the owner's call once the project resumes.

The documents are consistent with that state. The plan's own row says 0 of 7 tasks done at 8ad14ca919 (WorkItem has no parent field yet), and "it changes no code" is true of this diff. CI is green on this head.

When the project resumes, please move the status line and both README rows off "on hold" in the same push that asks for review. The sign-off can then be judged against the current design.

@github-actions github-actions Bot added the merge conflict Branch has merge conflicts with its base — author must resolve before merge label Nov 15, 2026
@chenmingwei23
chenmingwei23 force-pushed the docs/rfc-ultra-mode-task-hierarchy branch from a1220e8 to 031039d Compare November 16, 2026 04:24
@chenmingwei23

Copy link
Copy Markdown
Contributor Author

Rebased onto main to clear a conflict, no content change.

@github-actions github-actions Bot added readiness: checking Automated validation is still running readiness: passed Eligible automated validation passed for the current revision and removed readiness: passed Eligible automated validation passed for the current revision merge conflict Branch has merge conflicts with its base — author must resolve before merge readiness: checking Automated validation is still running labels Nov 16, 2026
@chenmingwei23

Copy link
Copy Markdown
Contributor Author

Closing. The project is on hold, and section 6 of this RFC is now out of date: it keeps a separate team-lead skill, while decision #37696 (merged 2026-11-15) moves conductor mode back to kirocrew-conductor and folds the useful team-lead rules into goal-conductor. The branch stays; tracking issue #34239 keeps the design link. If the task hierarchy work resumes, reopen and rewrite section 6 against #37696 first.

@github-actions github-actions Bot removed the readiness: passed Eligible automated validation passed for the current revision label Nov 16, 2026
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.

3 participants