From b4d0197c75a286fb8fe02860b9a0a53a96fb1c7b Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 4 Oct 2026 11:10:14 +0000 Subject: [PATCH 1/3] spec(epic): FIX-1649 wrap amendment, the path as shipped and leg c as the closure checks it Redraw figures/path.svg from the merged PRs, 30 Sep to 4 Oct: every child that landed, the v2-look slices of FIX-1737 and the late fixes FIX-1730..1736, 1747, 1749, with the closure's runs and the critical path through design v2. Reconcile leg c in SPEC, BUSINESS-RULES (ER-13) and PLAN with FIX-1663's goal check: the sweep is the registry cards Shift Manager draws plus the devtool page (c:sweep), hardcoded-accent fails at c:tool alone, worker-session may reach c:sweep, and the static half is c:static. EVOLUTION records the closure's two calls: Inbox's workstreams per BR-18 (a1 reads its set from the store) and ER-15's room-kind allowance for sendAction( in lib/talk.ts and createSession( in lib/reads.ts, already listed in ER-15 and not repeated. Co-Authored-By: Claude Opus 5.5 (1M context) Claude-Session: https://claude.ai/code/session_01Cnm5DbhFmkarpesBpPQ9Dq --- specs/epics/FIX-1649/BUSINESS-RULES.md | 4 +- specs/epics/FIX-1649/EVOLUTION.md | 21 +++ specs/epics/FIX-1649/PLAN.md | 21 ++- specs/epics/FIX-1649/SPEC.md | 29 ++-- specs/epics/FIX-1649/figures/path.svg | 186 ++++++++++++++----------- 5 files changed, 155 insertions(+), 106 deletions(-) diff --git a/specs/epics/FIX-1649/BUSINESS-RULES.md b/specs/epics/FIX-1649/BUSINESS-RULES.md index 2b06ad3bf4..852fae89fc 100644 --- a/specs/epics/FIX-1649/BUSINESS-RULES.md +++ b/specs/epics/FIX-1649/BUSINESS-RULES.md @@ -23,7 +23,7 @@ checks. Each says who owns it and where it's checked. | ER-6 | No child restyles an FSD component or hardcodes paint on it inside App Lab. Registry components arrive by `fsdev ui add` and stay unedited copies of their source; a component that can't take the skin is fixed at its source by FIX-1655 and re-synced, and FIX-1655's re-sync check fails on any App Lab copy that differs from its source | [D2](DECISIONS.md#d2). A restyled copy drifts from every later FSD fix, and a stale one makes leg c test the copy, not the skin | | ER-7 | No child adds an L1 noun (Agent, Team, Channel, MessageBoard, Project-as-required) or puts Workforce concepts into `core` or `engine`; the design-system package imports nothing from `@flow-state-dev/workforce` | Jake's layer rule. L1 offers generic hooks only | | ER-8 | No child builds anything below [the box's fence](SPEC.md#whats-in-the-box), and none adds `CHANNELS.md` or `kind:` frontmatter | The PRD's and the Architect's invent-kills, and D-12; the fence is their one list | -| ER-15 | No child writes into a worker's session except through that session's shipped operations: a turn from the task composer, a workstream's `@worker` or a reply from Inbox, Interrupt, Hand off. The shell writes no session item of its own and shows a message as delivered only when the session says so. Owned by FIX-1664; FIX-1662's composers (a workstream's, Inbox's reply) consume it. Two writes go to the person's own sessions, never a worker's:
· a project room's `read`, `post` and `join`, on the room kind (FIX-1718)
· opening the person's room session, on the room kind (FIX-1718, FIX-1752) | A second write path into a harness session is a model the shell invented; the stream's "sent into PAY-14" line must be the session's answer, not the shell's hope | +| ER-15 | No child writes into a worker's session except through that session's shipped operations: a turn from the task composer, a workstream's `@worker` or a reply from Inbox, Interrupt, Hand off. The shell writes no session item of its own and shows a message as delivered only when the session says so. Owned by FIX-1664; FIX-1662's composers (a workstream's, Inbox's reply) consume it. Two writes go to the person's own sessions, never a worker's:
· a project room's `read`, `post` and `join`, on the room kind (FIX-1718)
· opening the person's room session, on the room kind (FIX-1718, FIX-1752)
Both decided at closure, 2026-10-04 ([EVOLUTION](EVOLUTION.md#amendment--2026-10-04--wrap)) | A second write path into a harness session is a model the shell invented; the stream's "sent into PAY-14" line must be the session's answer, not the shell's hope | ## How the set is run @@ -38,5 +38,5 @@ checks. Each says who owns it and where it's checked. | # | The epic is done when | Proved by | |---|---|---| -| ER-13 | [The goal](SPEC.md#the-goal-and-how-well-know-its-met) is met: legs a, b, c and d pass, leg c fails under its control, and leg d under each of FIX-1737's three controls (leg d added 2026-10-02) | FIX-1663's goal check, in a browser, real model | +| ER-13 | [The goal](SPEC.md#the-goal-and-how-well-know-its-met) is met: legs a, b, c and d pass, leg c fails at `c:tool` alone under `hardcoded-accent` (FIX-1664's `worker-session` may also reach `c:sweep`, and nothing else reaches leg c), and leg d under each of FIX-1737's three controls (leg d added 2026-10-02; leg c's signals named at wrap, 2026-10-04) | FIX-1663's goal check, in a browser, real model | | ER-14 | The docs teach skinning FSD components from one token set, and App Lab's README says how to open a Lab in it, what each level and destination shows, and how to switch day and night | FIX-1655's, FIX-1662's, FIX-1664's, FIX-1722's, FIX-1723's and FIX-1725's docs, per [DOCS.md](DOCS.md) | diff --git a/specs/epics/FIX-1649/EVOLUTION.md b/specs/epics/FIX-1649/EVOLUTION.md index 05dd4a07f1..40ee37c6a5 100644 --- a/specs/epics/FIX-1649/EVOLUTION.md +++ b/specs/epics/FIX-1649/EVOLUTION.md @@ -58,3 +58,24 @@ Out of this amendment, and staying with their owners: rows waiting on FIX-1650, data (54 rows, some only in part), FIX-1675's watches and FIX-1474's also-post, and the v1 structure the 2026-10-01 amendment kept (X1 to X7). FIX-1736 and FIX-1737 block FIX-1663 in Linear. + + +## Amendment · 2026-10-04 · wrap + +The closure, FIX-1663, merged on [#2709](https://github.com/fixpoint-labs/flow-state-dev/pull/2709) +(`894c418d2`) after a clean run on `01444863c`: every leg and part passed, and every control was +right. Its goal check, +[`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) with +`legs.mts` beside it, is the source of truth for what leg c checks. Two calls were made during the +closure; they are recorded here so no later Lab reopens them. The review records stay +[#2421](https://github.com/fixpoint-labs/flow-state-dev/pull/2421) and the two amendments above. + +| Prior intent in this set | Treatment | Reason / evidence | Replacement | Compatibility | +|---|---|---|---|---| +| *"Leg c's sweep, pinned once"*: the react chrome App Lab mounts (navigator, roster, board panels, seat detail), every registry component App Lab copies, the devtool page ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, narrowed to what Shift Manager draws | Shift Manager mounts no `@flow-state-dev/react` chrome; it imports providers, hooks and the item renderer from it, no chrome component. FIX-1663 narrowed its own plan after closure run 1 at `899e059` (EM call with FIX-1688 and FIX-1689, [#2532](https://github.com/fixpoint-labs/flow-state-dev/pull/2532)) and left this text to the wrap ([FIX-1663's evolution](../../issues/FIX-1663/EVOLUTION.md)) | The message, reasoning, tool, code block and ask cards, and the devtool page the trace link opens, on a build with only the theme import removed, day then night. Each part renders at least once per pass (`c:sweep`); a Shift Manager value on a part fails at `c:`; FIX-1655's static check over `packages/**` is `c:static` | ER-2, ER-3 and the theme's value list are unchanged. [D2](DECISIONS.md#d2)'s two shelves stay the contract a host may use; Shift Manager uses the registry shelf only | +| Leg c's control: *"one reused component given a hardcoded App Lab colour: leg c must FAIL naming it"* ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met), ER-13) | **Amended**, named | `hardcoded-accent` patches the tool card's copy; on `01444863c` it failed at `c:tool` only. FIX-1664's `worker-session` points the run-lab's Session, where leg c draws four of its parts, elsewhere, so goal.md lets it turn `c:sweep` red beside its `a2` | `hardcoded-accent` must fail at `c:tool` alone; `worker-session` may also reach `c:sweep`; no other control reaches leg c | The control's patch and the theme values are unchanged | +| Leg a's Inbox journey: the ask *"in Inbox with its kind and workstream"*, its card *"in the workstream's stream"*, read as the asking seat's first channel ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, the reading only | Once FIX-1718 put `eng.em` in four channels, Inbox named all four for an ask with no workstream session, and the specs left that case open (the run on `dd9054581` was red at a1 alone). Decided for the closure on 2026-10-04: such an ask belongs to every channel the store lists its seat in, because FIX-1662's [BR-18](../../issues/FIX-1662/BUSINESS-RULES.md) draws its card on each of those Streams; an ask a channel post started belongs to that channel ([`6c6030939`](https://github.com/fixpoint-labs/flow-state-dev/commit/6c6030939)) | a1 reads its expected set from the store, holds Inbox to exactly that set, and checks the card on each of those Streams. On `25d026528` and `01444863c`: eng.feature, eng.triage, ops.oncall, ops.release | BR-18 and Inbox are unchanged: the shipped behaviour was right and the check was too narrow | +| ER-15: no child writes into a worker's session except through its shipped operations ([BUSINESS-RULES](BUSINESS-RULES.md#what-no-child-may-do)) | **Amended**, narrowly | FIX-1718 and FIX-1752 added two writes part 4 did not name: `sendAction(` in `labs/shift-manager/src/lib/talk.ts` and `createSession(` in `lib/reads.ts`. Both write the person's own sessions on `ROOM_KIND`, never a worker's. Decided for the closure on 2026-10-04 ([`c59c57491`](https://github.com/fixpoint-labs/flow-state-dev/commit/c59c57491)) | Part 4 allows `sendAction(` in `lib/talk.ts` only for a project room's `read`, `post` and `join` on `ROOM_KIND`, and `createSession(` in `lib/reads.ts` only on `ROOM_KIND`. ER-15 already lists both writes; they are not repeated. On `6e0201344` each negative (a `hire` talk action, a non-room `joinRoom` or `createSession`, a second `sendAction`) turned part 4 red at `P4:ER-15` alone | No new write into a worker's session; ER-15's other doors and its owner are unchanged | + +[The path](PLAN.md#the-path) is redrawn as it shipped. The set table stays the 30 September +snapshot; live state is Linear and the implementation PRs. diff --git a/specs/epics/FIX-1649/PLAN.md b/specs/epics/FIX-1649/PLAN.md index 1a8db29d52..a53b98b3ab 100644 --- a/specs/epics/FIX-1649/PLAN.md +++ b/specs/epics/FIX-1649/PLAN.md @@ -8,19 +8,14 @@ and [BUSINESS-RULES.md](BUSINESS-RULES.md) (ER-n). ## The path -![The path, as of 30 September 2026, drawn in phases rather than dates because Cycle PM has not scheduled the build. An input lane for the design: hand-back v1 came back during review; design pass 2 runs from the gate to the final hand-back, a vertical line that gates final visuals only. An input lane for the sibling epics, which nothing waits on. This epic spec in review, with the now line on it. FIX-1655: spec, then the token contract, neutral defaults, skin fixes and the re-sync check, then the light and dark App Lab values after the final hand-back. FIX-1662: spec, then the frame and the project and workstream levels, then its final visual pass. FIX-1664: spec beside FIX-1662's, then the task level once FIX-1662's frame lands, then its final visual pass. FIX-1663: the QA plan written beside them, and its run after all three merge. The critical path runs the gate, design pass 2, the final passes and the closure run](figures/path.svg) - -Three lanes run from the gate, and all split at the same wait: the final hand-back. Everything -structural starts at the gate, except the task view's build, which waits for FIX-1662's frame; -only the final look waits on design ([ER-9](BUSINESS-RULES.md#how-the-set-is-run)). So design -pass 2, not any child, is the critical path's long pole. The dependency graph is in -[the spec](SPEC.md#how-the-issues-flow-into-each-other); this adds time. - -The figure is the 30 September path. The 2026-10-01 amendment adds three lanes after it: FIX-1722 -and FIX-1723 (spec, then build in the shipped frame) and FIX-1725 (built directly). They start -after the final hand-back, and the closure run now waits for them too. The 2026-10-02 amendment -adds two more: FIX-1736 (the fonts, built directly), then FIX-1737 (v2's look: spec, then four -PRs), and the closure's leg d runs FIX-1737's check. +![The path as it shipped, 30 September to 4 October 2026, one lane per issue or group, each bar from filed or the gate to merged. FIX-1655, FIX-1662 and FIX-1664, with the DevForce lab prerequisites FIX-1666 to 1668, specced and built on 30 September. Closure run 1's findings, FIX-1684 and FIX-1688 to 1696, merged overnight, and FIX-1690's door into a running session on 1 October. Design v2 landed that evening, followed by FIX-1725, FIX-1722 and FIX-1723. The late fixes FIX-1730 to 1735, FIX-1747 and FIX-1749 landed on 2 and 3 October. The v2-look audit filed FIX-1736, merged 2 October, and FIX-1737, whose four slices landed early on 4 October beside FIX-1650's FIX-1718 and FIX-1752 as inputs. The closure ran until one run was clean, on 01444863c, and PR 2709 merged on 4 October. The critical path is outlined: design v2, FIX-1736, FIX-1737 and the clean closure run](figures/path.svg) + +The long pole was v2's look, not the shell: the structure the 30 September plan named merged on +the first day, and the last three went to design v2, the look it set +([ER-16](BUSINESS-RULES.md#what-a-team-gets-and-what-it-doesnt)) and the closure's findings. The +plan's critical path ran through design pass 2 and the final passes; that held, with FIX-1736 and +FIX-1737 as the final pass. Four children filed under the +epic (FIX-1671, 1673, 1675, 1705) are not on the path and did not block the closure. ## The design hand-back is an input, not a child diff --git a/specs/epics/FIX-1649/SPEC.md b/specs/epics/FIX-1649/SPEC.md index f33d8c8528..4bae0fcab5 100644 --- a/specs/epics/FIX-1649/SPEC.md +++ b/specs/epics/FIX-1649/SPEC.md @@ -28,6 +28,13 @@ screen with v2, which the closure runs as leg d. Rows waiting on FIX-1650 to 165 structure kept above, and the registry parts stay out. What moved: [EVOLUTION.md](EVOLUTION.md#amendment--2026-10-02--v2s-look). +**Amended 2026-10-04, at wrap: leg c as the closure checks it.** The closure (FIX-1663, +[#2709](https://github.com/fixpoint-labs/flow-state-dev/pull/2709)) passed on `01444863c` with +every control right. Leg c's sweep and control below now say what its goal check runs, and the +calls made during closure (Inbox's workstreams, ER-15's room writes) are recorded: +[EVOLUTION.md](EVOLUTION.md#amendment--2026-10-04--wrap). [The path](PLAN.md#the-path) is +redrawn as it shipped. + ## Four teams, before and after | A team that… | Today | After this epic | @@ -64,8 +71,8 @@ flowchart LR L2 --> P L3 -->|"no App Lab value on any swept part"| P L4 -->|"every element matches its v2 row"| P - C["control · one reused part with a hardcoded accent"] -.-> L3 - L3 -.->|"under the control"| F["must FAIL · names the part"] + C["control · the tool card's copy with a hardcoded accent"] -.-> L3 + L3 -.->|"under the control"| F["must FAIL at c:tool, and only there"] ``` Leg c makes "not baked in" checkable; its control plants the failure it exists to catch. Leg d @@ -74,15 +81,19 @@ makes "v2's look" checkable, with FIX-1737's controls. | How we verify | | |---|---| | **Goal check** | The closure issue's goal check ([FIX-1663](https://linear.app/fixpoint-labs/issue/FIX-1663)), in a browser, on one `main` commit after every other child merges ([ER-12](BUSINESS-RULES.md#how-the-set-is-run)) | -| **Signal** | Leg a lands on Chief of Staff, App Lab's first screen, and walks the design's two journeys from there: **Inbox → an approval → Approve & run**, and it leaves Inbox and its card in the workstream's stream together; **Tasks or the project board → a task → its Session**, with the task inspector beside it. Then every sidebar section, every tab at the three levels, the right panel at a workstream and at a task, and resources from Jump to are reached; a message to `@worker` in a workstream's composer arrives in that worker's session as a turn. Where FIX-1650 hasn't shipped projects, the project level passes on its named empty state and the second journey starts from Tasks or the workstream's Board. Leg b: the second tree opens through FIX-1662's load path and reaches the same surfaces, with no shell code of its own. Leg c: with no App Lab theme loaded, computed styles on every swept part carry no App Lab value, and a static check finds none under `packages/`. Leg d (amended 2026-10-02): FIX-1737's goal check (`goals/shift-manager/it-draws-v2s-look/`) on the same commit over the DevForce tree and its own keyless desk Lab, day and night: both fonts loaded, every look-table row rendered, and every visible element matches its row in a v2 look table that cites v2's lines, outside the rows that wait on a sibling and the registry parts it lists | +| **Signal** | Leg a lands on Chief of Staff, App Lab's first screen, and walks the design's two journeys from there: **Inbox → an approval → Approve & run**, and it leaves Inbox and its card on every workstream's Stream that draws it together (the channel a post started it from, else every channel the store lists the asking seat in, per FIX-1662's BR-18; decided at closure, [EVOLUTION](EVOLUTION.md#amendment--2026-10-04--wrap)); **Tasks or the project board → a task → its Session**, with the task inspector beside it. Then every sidebar section, every tab at the three levels, the right panel at a workstream and at a task, and resources from Jump to are reached; a message to `@worker` in a workstream's composer arrives in that worker's session as a turn. Where FIX-1650 hasn't shipped projects, the project level passes on its named empty state and the second journey starts from Tasks or the workstream's Board. Leg b: the second tree opens through FIX-1662's load path and reaches the same surfaces, with no shell code of its own. Leg c: on a build with only the theme import removed, day then night, the computed colour, background, border and font of every swept part match no value the Shift Manager theme declares (a hit fails at `c:`), and every swept part renders at least once per pass (`c:sweep`); FIX-1655's static check over `packages/**` finds none (`c:static`). Leg d (amended 2026-10-02): FIX-1737's goal check (`goals/shift-manager/it-draws-v2s-look/`) on the same commit over the DevForce tree and its own keyless desk Lab, day and night: both fonts loaded, every look-table row rendered, and every visible element matches its row in a v2 look table that cites v2's lines, outside the rows that wait on a sibling and the registry parts it lists | | **Input** | The DevForce lab's tree (`goals/devforce-lab/lab/workforce/`) with a real model, under a real org; for leg b, the pentest lab's tree (`goals/pentest-lab/lab/workforce/`), the nearest to CyberForce in the repo. Kitchen-sink's tree stands in for neither | | **Anti-game** | No asserting on a child's own tests. No App Lab code that names the second tree. No swept part left out, and no copy out of sync with its source | -| **Control that must fail** | One reused component given a hardcoded App Lab colour: leg c must FAIL naming it. FIX-1737's three controls, each must FAIL leg d where it names: `drift` (one row rounded, one ID column in sans), `unclassified` (an element no row covers), `missing` (the Tasks ID column removed). Today's `main`: legs a, b and d fail | - -**Leg c's sweep, pinned once:** the `@flow-state-dev/react` chrome App Lab mounts (navigator, -roster, board panels, seat detail), every `@flow-state-dev/ui` registry component App Lab copies -in (the stream's cards, tool calls and diffs among them), and the devtool page the trace link -opens. FIX-1663's QA plan lists them by name. +| **Control that must fail** | `hardcoded-accent`, the tool card's copy given Shift Manager's accent as a literal: leg c must FAIL at `c:tool` and nowhere else. FIX-1737's three controls, each must FAIL leg d where it names: `drift` (one row rounded, one ID column in sans), `unclassified` (an element no row covers), `missing` (the Tasks ID column removed). Today's `main`: legs a, b and d fail, and leg c's static half stays green | + +**Leg c's sweep, pinned once** (amended at wrap, [EVOLUTION](EVOLUTION.md#amendment--2026-10-04--wrap)): +the `@flow-state-dev/ui` registry cards Shift Manager draws, each found on the page that draws it: +the message, reasoning, tool, code block and ask cards (the ask in Inbox's detail and on the Stream), +and the devtool page the task inspector's trace link opens. Shift Manager mounts no +`@flow-state-dev/react` chrome, so no navigator, roster, board panel or seat detail is swept. A +pass that never draws a part fails at `c:sweep`; FIX-1664's `worker-session` control, which +points the run-lab's Session elsewhere, may turn `c:sweep` red beside its own `a2`; no control other +than these two may reach leg c. ## What's in the box diff --git a/specs/epics/FIX-1649/figures/path.svg b/specs/epics/FIX-1649/figures/path.svg index 7db4037f9f..ccf6056f69 100644 --- a/specs/epics/FIX-1649/figures/path.svg +++ b/specs/epics/FIX-1649/figures/path.svg @@ -1,5 +1,5 @@ - + - + - THE PATH — PHASES, NOT DATES · CYCLE 2 CANDIDATE, NOT SCHEDULED - - now · review - gate - build, in parallel - final hand-back - final visuals - closure run - - - - - Claude Design · input - - v1 · in - - design pass 2 → the final hand-back, with Jake - - - FIX-1650–1652 · input - meaning for the empty states · nothing here waits on it - - - FIX-1649 · this spec - - in review - - - FIX-1655 · design system - - spec - - contract · defaults · fixes · re-sync - - light, dark values + THE PATH AS IT SHIPPED · 30 SEP – 4 OCT 2026 · FILED (OR THE GATE) TO MERGED + + + Sep 30Oct 1 + Oct 2Oct 3 + Oct 4 + + v2 + now - - FIX-1662 · App Lab shell - - spec - - frame · two levels - - final visual pass - - - FIX-1664 · task view - - spec - - task level - - final visual pass - - - - FIX-1663 · closure - - QA plan - - run, one commit + + Claude Design · input + v1 in + + pass 2 → v2 · #2605 + FIX-1650 epic · input + FIX-1718 · FIX-1752 + + + FIX-1649 · epic spec + merged + owners + + rename · v2 + v2 look + wrap + FIX-1655 · design system + + spec · build + FIX-1662 · the shell + + frame · two levels + FIX-1664 · task view + + task level + FIX-1666–1668 · DevForce lab + + ask · board · run link + FIX-1684, 1688–96 · run 1 + + skin · trace + FIX-1690 · @worker turn + + door · live run + FIX-1725 · Day/Night + direct + FIX-1722 · Chief of Staff + + the landing view + FIX-1723 · Roster + + on shift · on call · off shift + FIX-1730–35, 1747, 1749 + + late fixes · 4 PRs + FIX-1736 · fonts + + direct + FIX-1737 · v2's look + + + specslices A, B, C, D · Oct 4 + FIX-1663 · closure + QA plan + + + + + + runs until clean + clean on 01444863c · #2709 ▸ + + + spec + build, merged + + closure run, red / clean + input from outside the set + the critical path - - Critical path: gate → design pass 2 → the final passes → the closure run. The task level waits on the frame, not on design. - Bars are phases, not durations; widths are not estimates. Linear and the implementation PRs hold live state. + Critical path, as shipped: design v2 → the v2-look audit → FIX-1736 → FIX-1737's spec and four slices → the clean closure run. + Filed under the epic but not on the path: FIX-1671, 1673, 1675, 1705. Linear and the implementation PRs hold live state. From daa7b79745776e8be61b669332b342b5f5caf9f1 Mon Sep 17 00:00:00 2001 From: Claude Date: Sun, 4 Oct 2026 11:14:43 +0000 Subject: [PATCH 2/3] spec(epic): FIX-1649 wrap, fold Cursor's review: goal.md is leg c's one home SPEC's Signal cell is one line per leg, with leg c linked to the closure's goal.md as leg d links to it-draws-v2s-look; the sweep block is one sentence; the today's-main clause is scoped to that control on a pre-Shift-Manager commit. ER-13 points at FIX-1663's goal check instead of naming c:* signals. EVOLUTION's wrap rows keep the why and end at the SPEC goal section plus goal.md; the ER-15 row links to the rule in BUSINESS-RULES. path.svg's aria-label and PLAN's alt are short; the timeline lives in one PLAN paragraph. Co-Authored-By: Claude Opus 5.5 (1M context) Claude-Session: https://claude.ai/code/session_01Cnm5DbhFmkarpesBpPQ9Dq --- specs/epics/FIX-1649/BUSINESS-RULES.md | 2 +- specs/epics/FIX-1649/EVOLUTION.md | 10 +++++----- specs/epics/FIX-1649/PLAN.md | 11 ++++++++++- specs/epics/FIX-1649/SPEC.md | 22 ++++++++-------------- specs/epics/FIX-1649/figures/path.svg | 2 +- 5 files changed, 25 insertions(+), 22 deletions(-) diff --git a/specs/epics/FIX-1649/BUSINESS-RULES.md b/specs/epics/FIX-1649/BUSINESS-RULES.md index 852fae89fc..9cc4c14dd5 100644 --- a/specs/epics/FIX-1649/BUSINESS-RULES.md +++ b/specs/epics/FIX-1649/BUSINESS-RULES.md @@ -38,5 +38,5 @@ checks. Each says who owns it and where it's checked. | # | The epic is done when | Proved by | |---|---|---| -| ER-13 | [The goal](SPEC.md#the-goal-and-how-well-know-its-met) is met: legs a, b, c and d pass, leg c fails at `c:tool` alone under `hardcoded-accent` (FIX-1664's `worker-session` may also reach `c:sweep`, and nothing else reaches leg c), and leg d under each of FIX-1737's three controls (leg d added 2026-10-02; leg c's signals named at wrap, 2026-10-04) | FIX-1663's goal check, in a browser, real model | +| ER-13 | [The goal](SPEC.md#the-goal-and-how-well-know-its-met) is met: legs a to d pass and each control fails where it must, per [FIX-1663's goal check](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) (leg d added 2026-10-02) | FIX-1663's goal check, in a browser, real model | | ER-14 | The docs teach skinning FSD components from one token set, and App Lab's README says how to open a Lab in it, what each level and destination shows, and how to switch day and night | FIX-1655's, FIX-1662's, FIX-1664's, FIX-1722's, FIX-1723's and FIX-1725's docs, per [DOCS.md](DOCS.md) | diff --git a/specs/epics/FIX-1649/EVOLUTION.md b/specs/epics/FIX-1649/EVOLUTION.md index 40ee37c6a5..8bf3fc3303 100644 --- a/specs/epics/FIX-1649/EVOLUTION.md +++ b/specs/epics/FIX-1649/EVOLUTION.md @@ -66,16 +66,16 @@ The closure, FIX-1663, merged on [#2709](https://github.com/fixpoint-labs/flow-s (`894c418d2`) after a clean run on `01444863c`: every leg and part passed, and every control was right. Its goal check, [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) with -`legs.mts` beside it, is the source of truth for what leg c checks. Two calls were made during the +`legs.mts` beside it, is the one home for how legs a to c are checked, as `it-draws-v2s-look/` is for leg d. Two calls were made during the closure; they are recorded here so no later Lab reopens them. The review records stay [#2421](https://github.com/fixpoint-labs/flow-state-dev/pull/2421) and the two amendments above. | Prior intent in this set | Treatment | Reason / evidence | Replacement | Compatibility | |---|---|---|---|---| -| *"Leg c's sweep, pinned once"*: the react chrome App Lab mounts (navigator, roster, board panels, seat detail), every registry component App Lab copies, the devtool page ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, narrowed to what Shift Manager draws | Shift Manager mounts no `@flow-state-dev/react` chrome; it imports providers, hooks and the item renderer from it, no chrome component. FIX-1663 narrowed its own plan after closure run 1 at `899e059` (EM call with FIX-1688 and FIX-1689, [#2532](https://github.com/fixpoint-labs/flow-state-dev/pull/2532)) and left this text to the wrap ([FIX-1663's evolution](../../issues/FIX-1663/EVOLUTION.md)) | The message, reasoning, tool, code block and ask cards, and the devtool page the trace link opens, on a build with only the theme import removed, day then night. Each part renders at least once per pass (`c:sweep`); a Shift Manager value on a part fails at `c:`; FIX-1655's static check over `packages/**` is `c:static` | ER-2, ER-3 and the theme's value list are unchanged. [D2](DECISIONS.md#d2)'s two shelves stay the contract a host may use; Shift Manager uses the registry shelf only | -| Leg c's control: *"one reused component given a hardcoded App Lab colour: leg c must FAIL naming it"* ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met), ER-13) | **Amended**, named | `hardcoded-accent` patches the tool card's copy; on `01444863c` it failed at `c:tool` only. FIX-1664's `worker-session` points the run-lab's Session, where leg c draws four of its parts, elsewhere, so goal.md lets it turn `c:sweep` red beside its `a2` | `hardcoded-accent` must fail at `c:tool` alone; `worker-session` may also reach `c:sweep`; no other control reaches leg c | The control's patch and the theme values are unchanged | -| Leg a's Inbox journey: the ask *"in Inbox with its kind and workstream"*, its card *"in the workstream's stream"*, read as the asking seat's first channel ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, the reading only | Once FIX-1718 put `eng.em` in four channels, Inbox named all four for an ask with no workstream session, and the specs left that case open (the run on `dd9054581` was red at a1 alone). Decided for the closure on 2026-10-04: such an ask belongs to every channel the store lists its seat in, because FIX-1662's [BR-18](../../issues/FIX-1662/BUSINESS-RULES.md) draws its card on each of those Streams; an ask a channel post started belongs to that channel ([`6c6030939`](https://github.com/fixpoint-labs/flow-state-dev/commit/6c6030939)) | a1 reads its expected set from the store, holds Inbox to exactly that set, and checks the card on each of those Streams. On `25d026528` and `01444863c`: eng.feature, eng.triage, ops.oncall, ops.release | BR-18 and Inbox are unchanged: the shipped behaviour was right and the check was too narrow | -| ER-15: no child writes into a worker's session except through its shipped operations ([BUSINESS-RULES](BUSINESS-RULES.md#what-no-child-may-do)) | **Amended**, narrowly | FIX-1718 and FIX-1752 added two writes part 4 did not name: `sendAction(` in `labs/shift-manager/src/lib/talk.ts` and `createSession(` in `lib/reads.ts`. Both write the person's own sessions on `ROOM_KIND`, never a worker's. Decided for the closure on 2026-10-04 ([`c59c57491`](https://github.com/fixpoint-labs/flow-state-dev/commit/c59c57491)) | Part 4 allows `sendAction(` in `lib/talk.ts` only for a project room's `read`, `post` and `join` on `ROOM_KIND`, and `createSession(` in `lib/reads.ts` only on `ROOM_KIND`. ER-15 already lists both writes; they are not repeated. On `6e0201344` each negative (a `hire` talk action, a non-room `joinRoom` or `createSession`, a second `sendAction`) turned part 4 red at `P4:ER-15` alone | No new write into a worker's session; ER-15's other doors and its owner are unchanged | +| *"Leg c's sweep, pinned once"*: the react chrome App Lab mounts (navigator, roster, board panels, seat detail), every registry component App Lab copies, the devtool page ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, narrowed to what Shift Manager draws | Shift Manager mounts no `@flow-state-dev/react` chrome; it imports providers, hooks and the item renderer from it, no chrome component. FIX-1663 narrowed its own plan after closure run 1 at `899e059` (EM call with FIX-1688 and FIX-1689, [#2532](https://github.com/fixpoint-labs/flow-state-dev/pull/2532)) and left this text to the wrap ([FIX-1663's evolution](../../issues/FIX-1663/EVOLUTION.md)) | Leg c sweeps the registry cards Shift Manager draws and the devtool page, not react chrome: [SPEC goal §](SPEC.md#the-goal-and-how-well-know-its-met) + [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) | ER-2, ER-3 and the theme's value list are unchanged. [D2](DECISIONS.md#d2)'s two shelves stay the contract a host may use; Shift Manager uses the registry shelf only | +| Leg c's control: *"one reused component given a hardcoded App Lab colour: leg c must FAIL naming it"* ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met), ER-13) | **Amended**, named | `hardcoded-accent` patches the tool card's copy; on `01444863c` it failed at `c:tool` only. FIX-1664's `worker-session` points the run-lab's Session, where leg c draws four of its parts, elsewhere, so goal.md lets it turn `c:sweep` red beside its `a2` | Each control fails at its own signal only, as the closure check grades it: [SPEC goal §](SPEC.md#the-goal-and-how-well-know-its-met) + [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) | The control's patch and the theme values are unchanged | +| Leg a's Inbox journey: the ask *"in Inbox with its kind and workstream"*, its card *"in the workstream's stream"*, read as the asking seat's first channel ([SPEC](SPEC.md#the-goal-and-how-well-know-its-met)) | **Amended**, the reading only | Once FIX-1718 put `eng.em` in four channels, Inbox named all four for an ask with no workstream session, and the specs left that case open (the run on `dd9054581` was red at a1 alone). Decided for the closure on 2026-10-04: such an ask belongs to every channel the store lists its seat in, because FIX-1662's [BR-18](../../issues/FIX-1662/BUSINESS-RULES.md) draws its card on each of those Streams; an ask a channel post started belongs to that channel ([`6c6030939`](https://github.com/fixpoint-labs/flow-state-dev/commit/6c6030939)) | a1 reads its expected workstreams from the store and holds Inbox to exactly that set: [SPEC goal §](SPEC.md#the-goal-and-how-well-know-its-met) + [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) | BR-18 and Inbox are unchanged: the shipped behaviour was right and the check was too narrow | +| ER-15: no child writes into a worker's session except through its shipped operations ([BUSINESS-RULES](BUSINESS-RULES.md#what-no-child-may-do)) | **Amended**, narrowly | FIX-1718 and FIX-1752 added two writes part 4 did not name: `sendAction(` in `labs/shift-manager/src/lib/talk.ts` and `createSession(` in `lib/reads.ts`. Both write the person's own sessions on `ROOM_KIND`, never a worker's. Decided for the closure on 2026-10-04 ([`c59c57491`](https://github.com/fixpoint-labs/flow-state-dev/commit/c59c57491)) | The two room-kind writes, as [ER-15](BUSINESS-RULES.md#what-no-child-may-do) states them; part 4 grades them: [SPEC goal §](SPEC.md#the-goal-and-how-well-know-its-met) + [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) | No new write into a worker's session; ER-15's other doors and its owner are unchanged | [The path](PLAN.md#the-path) is redrawn as it shipped. The set table stays the 30 September snapshot; live state is Linear and the implementation PRs. diff --git a/specs/epics/FIX-1649/PLAN.md b/specs/epics/FIX-1649/PLAN.md index a53b98b3ab..d57074b51b 100644 --- a/specs/epics/FIX-1649/PLAN.md +++ b/specs/epics/FIX-1649/PLAN.md @@ -8,7 +8,16 @@ and [BUSINESS-RULES.md](BUSINESS-RULES.md) (ER-n). ## The path -![The path as it shipped, 30 September to 4 October 2026, one lane per issue or group, each bar from filed or the gate to merged. FIX-1655, FIX-1662 and FIX-1664, with the DevForce lab prerequisites FIX-1666 to 1668, specced and built on 30 September. Closure run 1's findings, FIX-1684 and FIX-1688 to 1696, merged overnight, and FIX-1690's door into a running session on 1 October. Design v2 landed that evening, followed by FIX-1725, FIX-1722 and FIX-1723. The late fixes FIX-1730 to 1735, FIX-1747 and FIX-1749 landed on 2 and 3 October. The v2-look audit filed FIX-1736, merged 2 October, and FIX-1737, whose four slices landed early on 4 October beside FIX-1650's FIX-1718 and FIX-1752 as inputs. The closure ran until one run was clean, on 01444863c, and PR 2709 merged on 4 October. The critical path is outlined: design v2, FIX-1736, FIX-1737 and the clean closure run](figures/path.svg) +![The as-shipped path, 30 September to 4 October 2026; the detail is in the prose below](figures/path.svg) + +Each bar runs from filed (or the gate) to merged. FIX-1655, FIX-1662 and FIX-1664, with the DevForce +lab prerequisites FIX-1666 to 1668, specced and built on 30 September. Closure run 1's findings +(FIX-1684, FIX-1688 to 1696) merged overnight, and FIX-1690's door into a running session on 1 +October. Design v2 landed that evening, then FIX-1725, FIX-1722 and FIX-1723. The late fixes, +FIX-1730 to 1735, FIX-1747 and FIX-1749, landed on 2 and 3 October. The v2-look audit filed +FIX-1736, merged on 2 October, and FIX-1737, whose four slices landed early on 4 October beside +FIX-1650's FIX-1718 and FIX-1752. The closure ran until one run was clean, on `01444863c`, and +[#2709](https://github.com/fixpoint-labs/flow-state-dev/pull/2709) merged that day. The long pole was v2's look, not the shell: the structure the 30 September plan named merged on the first day, and the last three went to design v2, the look it set diff --git a/specs/epics/FIX-1649/SPEC.md b/specs/epics/FIX-1649/SPEC.md index 4bae0fcab5..0c253ec1eb 100644 --- a/specs/epics/FIX-1649/SPEC.md +++ b/specs/epics/FIX-1649/SPEC.md @@ -30,7 +30,7 @@ structure kept above, and the registry parts stay out. What moved: **Amended 2026-10-04, at wrap: leg c as the closure checks it.** The closure (FIX-1663, [#2709](https://github.com/fixpoint-labs/flow-state-dev/pull/2709)) passed on `01444863c` with -every control right. Leg c's sweep and control below now say what its goal check runs, and the +every control right. Legs a to c now point at its goal check for their mechanics, as leg d does, and the calls made during closure (Inbox's workstreams, ER-15's room writes) are recorded: [EVOLUTION.md](EVOLUTION.md#amendment--2026-10-04--wrap). [The path](PLAN.md#the-path) is redrawn as it shipped. @@ -71,8 +71,8 @@ flowchart LR L2 --> P L3 -->|"no App Lab value on any swept part"| P L4 -->|"every element matches its v2 row"| P - C["control · the tool card's copy with a hardcoded accent"] -.-> L3 - L3 -.->|"under the control"| F["must FAIL at c:tool, and only there"] + C["control · a reused card's copy with a hardcoded accent"] -.-> L3 + L3 -.->|"under the control"| F["must FAIL at that card, and only there"] ``` Leg c makes "not baked in" checkable; its control plants the failure it exists to catch. Leg d @@ -81,19 +81,13 @@ makes "v2's look" checkable, with FIX-1737's controls. | How we verify | | |---|---| | **Goal check** | The closure issue's goal check ([FIX-1663](https://linear.app/fixpoint-labs/issue/FIX-1663)), in a browser, on one `main` commit after every other child merges ([ER-12](BUSINESS-RULES.md#how-the-set-is-run)) | -| **Signal** | Leg a lands on Chief of Staff, App Lab's first screen, and walks the design's two journeys from there: **Inbox → an approval → Approve & run**, and it leaves Inbox and its card on every workstream's Stream that draws it together (the channel a post started it from, else every channel the store lists the asking seat in, per FIX-1662's BR-18; decided at closure, [EVOLUTION](EVOLUTION.md#amendment--2026-10-04--wrap)); **Tasks or the project board → a task → its Session**, with the task inspector beside it. Then every sidebar section, every tab at the three levels, the right panel at a workstream and at a task, and resources from Jump to are reached; a message to `@worker` in a workstream's composer arrives in that worker's session as a turn. Where FIX-1650 hasn't shipped projects, the project level passes on its named empty state and the second journey starts from Tasks or the workstream's Board. Leg b: the second tree opens through FIX-1662's load path and reaches the same surfaces, with no shell code of its own. Leg c: on a build with only the theme import removed, day then night, the computed colour, background, border and font of every swept part match no value the Shift Manager theme declares (a hit fails at `c:`), and every swept part renders at least once per pass (`c:sweep`); FIX-1655's static check over `packages/**` finds none (`c:static`). Leg d (amended 2026-10-02): FIX-1737's goal check (`goals/shift-manager/it-draws-v2s-look/`) on the same commit over the DevForce tree and its own keyless desk Lab, day and night: both fonts loaded, every look-table row rendered, and every visible element matches its row in a v2 look table that cites v2's lines, outside the rows that wait on a sibling and the registry parts it lists | +| **Signal** | One line per leg; the mechanics live in the closure's [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) and its `legs.mts`.
· **Leg a:** from Chief of Staff, the design's two journeys (Inbox → an approval → Approve & run; Tasks or the Board → a task → its Session), every surface the shell promises, and an `@worker` turn arriving in that worker's session. Which workstreams an ask belongs to is read per FIX-1662's BR-18 ([decided at closure](EVOLUTION.md#amendment--2026-10-04--wrap))
· **Leg b:** the second tree reaches the same surfaces through FIX-1662's load path, with no shell code of its own
· **Leg c:** with no Shift Manager theme loaded, no swept part carries a theme value, day or night, and none sits under `packages/` ([`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md))
· **Leg d:** FIX-1737's goal check, [`goals/shift-manager/it-draws-v2s-look/`](../../../goals/shift-manager/it-draws-v2s-look/goal.md), on the same commit | | **Input** | The DevForce lab's tree (`goals/devforce-lab/lab/workforce/`) with a real model, under a real org; for leg b, the pentest lab's tree (`goals/pentest-lab/lab/workforce/`), the nearest to CyberForce in the repo. Kitchen-sink's tree stands in for neither | | **Anti-game** | No asserting on a child's own tests. No App Lab code that names the second tree. No swept part left out, and no copy out of sync with its source | -| **Control that must fail** | `hardcoded-accent`, the tool card's copy given Shift Manager's accent as a literal: leg c must FAIL at `c:tool` and nowhere else. FIX-1737's three controls, each must FAIL leg d where it names: `drift` (one row rounded, one ID column in sans), `unclassified` (an element no row covers), `missing` (the Tasks ID column removed). Today's `main`: legs a, b and d fail, and leg c's static half stays green | - -**Leg c's sweep, pinned once** (amended at wrap, [EVOLUTION](EVOLUTION.md#amendment--2026-10-04--wrap)): -the `@flow-state-dev/ui` registry cards Shift Manager draws, each found on the page that draws it: -the message, reasoning, tool, code block and ask cards (the ask in Inbox's detail and on the Stream), -and the devtool page the task inspector's trace link opens. Shift Manager mounts no -`@flow-state-dev/react` chrome, so no navigator, roster, board panel or seat detail is swept. A -pass that never draws a part fails at `c:sweep`; FIX-1664's `worker-session` control, which -points the run-lab's Session elsewhere, may turn `c:sweep` red beside its own `a2`; no control other -than these two may reach leg c. +| **Control that must fail** | Leg c's `hardcoded-accent`, a literal accent on a reused card's copy, and leg d's `drift`, `unclassified` and `missing`, each failing where its goal check names. On a commit before Shift Manager's first merge, the `today's main` control fails legs a and b. The full list: [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md) | + +**Leg c's sweep** is what Shift Manager draws: the registry cards and the devtool page, not react +chrome it never mounts ([amended at wrap](EVOLUTION.md#amendment--2026-10-04--wrap); the parts by name in [`goal.md`](../../../goals/shift-manager/a-lab-is-worked-through-one-skinned-shell/goal.md)). ## What's in the box diff --git a/specs/epics/FIX-1649/figures/path.svg b/specs/epics/FIX-1649/figures/path.svg index ccf6056f69..161563ab82 100644 --- a/specs/epics/FIX-1649/figures/path.svg +++ b/specs/epics/FIX-1649/figures/path.svg @@ -1,5 +1,5 @@ + aria-label="The as-shipped path, 30 September to 4 October 2026, one lane per issue from filed to merged, with the critical path outlined; the detail is in PLAN.md, the path section.">