Repository navigation
spec(epic): record FIX-1793's and FIX-1791's gates (FIX-1786) - #2824
Conversation
The shared half stays in the MVP, and only a shared project's members open workstreams (Jake, 2026-10-06). Project claims and a project row's mailbox list move from FIX-1793 to FIX-1792, which removes them as it converts each board. ER-24: this record binds the set. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XMg2dz8sMEdyzCQTRsaMTV
|
The latest updates on your projects. Learn more about Vercel for GitHub. 3 Skipped Deployments
|
Jake answered FIX-1791 Q1 on 2026-10-06: filing tasks for delegates and following them through move to FIX-1794, and a coordinator files no task until FIX-1794 ships. ER-24: this record binds the set. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XMg2dz8sMEdyzCQTRsaMTV
…X-1791 Q1 and EVOLUTION) Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01XMg2dz8sMEdyzCQTRsaMTV
Codex Review SummaryThis comment shows the latest Codex review activity on this pull request.
ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings. |
There was a problem hiding this comment.
Simplify review (spec amend #5)
Problem this PR solves: Child gates on FIX-1793 and FIX-1791 produced evidence that binds the epic set only after an ER-24 follow-up on FIX-1786. This PR closes the forward-looking Q2 language and records ownership splits (claims → FIX-1792, task filing → FIX-1794, leg d → FIX-1793).
Is the approach too heavy? For this layer, mostly no. ER-24 explicitly requires a follow-up epic PR rather than treating the child spec merge as canonical for the set. Amendments 1–4 already established the same rhythm; a fifth PR is process cost, not accidental scope creep.
Cleaner alternative (process, not a blocker here): The duplication and drift come from splitting gate outcome across three surfaces:
- merged child spec (still lists Open: Q1/Q2 on FIX-1793 and Open: Q1 on FIX-1791),
- epic Q cards + Decided in review + How it got here,
- SPEC/PLAN status columns and coordination seams.
A simpler long-term pattern (worth considering for the next gate, not re-deriving this PR): when a child spec merges at gate, close the Q cards in that same PR (or a stacked child amend merged the same day), and let the epic follow-up be a thin ER-24 bind: update Q2 open/closed, one chronology bullet, and pointers—without re-pasting paragraphs that already live in the child’s decided-not-asked section. This PR re-states FIX-1792/claims and FIX-1794/task filing in four places; one canonical bullet under Decided in review plus links would be easier to keep aligned.
What looks right
- Replacing “ask at FIX-1793’s gate” with past tense in SPEC, PLAN step 8, and Open: none is the right hygiene.
- EVOLUTION compatibility row now matches FIX-1793 (deprecated claims, FIX-1792 removes)—fixes a real cross-child inconsistency.
- Leg d/e routing into FIX-1793 and FIX-1794 matches FIX-1791 Q1’s recommendation and tightens FIX-1791’s scope in PLAN/SPEC without reopening the model.
Follow-ups I’d treat as optional polish (not blocking merge)
- ER-24 in
BUSINESS-RULES.mdstill only says Q1 was met after FIX-1789’s gate; extending that sentence when Q2’s gate questions are fully recorded would help future readers. - Child specs still show open gate questions while the epic records answers—expect a FIX-1793/FIX-1791 amend pass so implementers aren’t reading two truths.
- Three commits could be one; purely cosmetic.
Verdict: Direction and content look correct for an epic ER-24 record. Approve; consider thinning future gate→epic amends and closing child Q cards at merge time to cut repetition.
Sent by Cursor Automation: Review For Best Approach
| private step and the docs phrase together. It didn't fire: FIX-1793's spec priced a private | ||
| project as today's row kept in user scope, a scope configuration (its D1). | ||
| - **The shared half stays in the MVP.** Jake, 2026-10-06, at FIX-1793's spec gate | ||
| ([#2823](https://github.com/fixpoint-labs/flow-state-dev/pull/2823)): "yes to all |
There was a problem hiding this comment.
Binding the whole gate to "yes to all recommendations" is fine for audit, but it bundles Q1 (shared half), Q2 (members open workstreams), and D1 (private = scope config) in one quote. If a future reader disputes one horn, they have to unpack the child spec. Optional simplify: keep the quote, add a one-line manifest right under it (shared in MVP · members-only open · private = user-scope row) so the epic card doesn’t rely on inferring all three from the blanket yes.
| From review round 2 (Codex). | ||
| - **Shift Manager lives at `packages/shift-manager`.** #2759 (FIX-1770) moves it there and lands | ||
| first among the merge-first rows; PRs that edit the lab rebase after it. | ||
| - **FIX-1792 removes project claims and a project row's mailbox list, not FIX-1793.** Today they |
There was a problem hiding this comment.
This bullet (and the FIX-1794 task-filing bullet below) largely mirror FIX-1793’s decided-not-asked / FIX-1791 Q1 prose. That’s defensible for ER-24 binding, but it’s the main complexity tax in this diff. Alternative for a future gate: epic amend adds a short pointer + ER-24 sentence here; keep the full mechanics only in the child spec unless the decision actually changes the epic goal or ER table.
| call 2. FIX-1774 is Canceled and FIX-1777 a Duplicate in Linear. FIX-1791 carries FIX-1774's | ||
| dogfood legs and its *not done if* list, and FIX-1794 carries FIX-1777's "runs as the filer" | ||
| rule, both restated on the new model. | ||
| rule, both restated on the new model. FIX-1774's leg e later moved to FIX-1794 and its leg d to FIX-1793 |
There was a problem hiding this comment.
Leg d/e now appear in Q3, Decided in review, How it got here, SPEC, and PLAN. Consider whether Q3 needs the leg split at all once the decided-in-review bullets exist—Q3 could stay “closed into FIX-1791/1794” and defer leg routing to the bullets below (minor dedup, not required for merge).
| | [FIX-1790](https://linear.app/fixpoint-labs/issue/FIX-1790) · user data per org | User-scoped data kept per (user, org), for every flow; records stored before moved to one org or refused | Without it a user's private workers show in every org they belong to | Backlog · spec route | | ||
| | [FIX-1788](https://linear.app/fixpoint-labs/issue/FIX-1788) · workers as resources | A worker as a user-scoped resource, run by the singleton flow it names; standard workers projected from files; fork; a session link callers can't seed; instances and pins deprecated | The spine: privacy by construction | Backlog · spec route | | ||
| | [FIX-1791](https://linear.app/fixpoint-labs/issue/FIX-1791) · coordinator flow | Delegates in session state, four routing policies, one answer per delegate, a routing record | Replaces the mailbox with a worker, and fixes its fixed membership | Backlog · spec route · carries FIX-1774's dogfood legs and *not done if* list | | ||
| | [FIX-1791](https://linear.app/fixpoint-labs/issue/FIX-1791) · coordinator flow | Delegates in session state, four routing policies, one answer per delegate, a routing record | Replaces the mailbox with a worker, and fixes its fixed membership | Backlog · spec route · carries FIX-1774's dogfood legs and *not done if* list, except leg d (FIX-1793's) and leg e (FIX-1794's) | |
There was a problem hiding this comment.
Status column is doing a lot of work (merged PR, Q2 outcomes, leg exceptions). That’s useful for the set snapshot, but it duplicates DECISIONS Q2/Q3. If the set table keeps growing gate paragraphs, a lighter pattern might be Spec merged (#2823) · gate recorded on epic [Q2](…) and leave the horns in DECISIONS only—soft suggestion, not a must-change.


Follow-up to #2795 (the FIX-1786 epic spec), amended by #2806, #2807, #2813 and #2822, all merged. It records two results of FIX-1793's spec gate (#2823) and one result of FIX-1791's spec gate. Per ER-24, a result from a child's gate binds the set only after the epic records it.
What changes
No rule in BUSINESS-RULES changes owner, because no rule covers claims or task filing. No figure changes, and the order of work does not change. The library (FIX-1795) remains after the MVP, as #2822 recorded.
Reviewers · look here
What's asked of you
Approve this record of your answers? I recommend yes. It writes down decisions you already made. Until this PR merges,
mainshows the shared half as open, FIX-1793 as the issue that removes claims, and FIX-1791 as the issue that files tasks.After merge (coordinator): no changes to Linear relations. FIX-1793 already blocks FIX-1792.
Links: FIX-1786 · original review #2795 · amendments #2806, #2807, #2813, #2822 · FIX-1793's spec #2823 · FIX-1791's spec amendment #2821 · set at
specs/epics/FIX-1786/How to review this: the epic-spec reviewer contract
This is an epic-spec: the shared objective and cross-cutting decisions for a set of
issues, in a retained document set. It is not any one issue's implementation design.
In scope to challenge:
SPEC.md— is this body of work worth doing, and is thegoal the real need at full size, or a smaller one the set could meet while the need stays
unmet? Would the closure issue's check fail if the goal were not met?
That question can only be asked here.
DECISIONS.md— shared surface, naming, sequencing, contracts.BUSINESS-RULES.mdwith no owner, or two.DOCS.md; missing or misleadingpredecessor treatment in conditional
EVOLUTION.md.Out of scope — owned by the individual issue specs:
poc/show shape and scoping, not shippable implementation. Directional evidence, experimental
isolation and absence of secrets/generated dependencies remain in scope.
Feedback in the second list is routed to the issue it concerns as an implementer note, not
folded in here.
Required repository checks, approvals and review-thread policy still govern merge.
Optional comments are triaged, not a reason to grind to zero or restart the design review.
🤖 Generated with Claude Code
https://claude.ai/code/session_01XMg2dz8sMEdyzCQTRsaMTV
Generated by Claude Code