Repository navigation
spec(FIX-2161): an ordinary session of any worker can open on a project - #3805
Conversation
Open the create check's project link to any worker at a random id, keep the project coordinator's derived-id rule, gate the coordinator flow's workstream seed on the project coordinator, and keep FIX-2158's list showing ordinary project sessions. Workstreams checked and not reused. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5
The create check for an ordinary project session asks the run's membership question, so a session can't open where projectWorkspace() refuses its runs. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5
|
The latest updates on your projects. Learn more about Vercel for GitHub. 3 Skipped Deployments
|
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. |
|
Coherence check (FSD Architect): the direction is coherent with epic FIX-2152 D3 (an ordinary session of one worker, no workstream per session, no new criteria key), and it adds no new mechanism: it reuses This spec and #3796 (FIX-2158) say opposite things about the same list. #3796 is approved and close to merging. Its D1 keeps only sessions whose criteria keys are a subset of Please close it in the spec text, not at implementation. Pick one:
Also, the body says "the epic thread has the conflict", but I don't see it on #3783. Please post it there or drop that line. Not approving the direction yet, only for this one point. Everything else is fine from a coherence view. — FSD Architect |
Second look + architecture pass (spec PR, checked against
|
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 9d8d9a5af2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
There was a problem hiding this comment.
Simplify review (spec FIX-2161)
Verdict: The implementation shape in this spec is about as small as it gets for the epic constraint (ordinary session of one worker, no workstream-per-session, no new criteria key). Re-reading installation.ts projectLinkCheck, filingProjectOf, and workstreamDelegateSeed on main supports the core claim: downstream readers already treat projectId as “where this session works”; the coordinator-only create refusal and the ungated delegate seed are the real blockers—not missing plumbing for filingProjectId or workstreams.
Alternatives considered (did the approach go too far?):
| Path | Simpler? | Why I’d still stick with this spec |
|---|---|---|
| Workstream per lab session | No | Epic D3 + 25-delegate cap + “project work” semantics |
New field (workProjectId) |
No | Fourth carrier; every reader doubles |
Lab passes input.project per task |
No | ER-4 glue; session’s own run has no project |
| Worker allow-list | No | Same check for every worker; doesn’t scale to next app |
Keep docs saying projectId = coordinator; only gate seed |
Partially | Same code as D2, but filing-session.ts and hand-over paths already document “works in”; keys.ts / projectLinkCheck comments are the outliers |
Reuse filingProjectId on create |
No | Pinned hand-over-only on main; would be a new writer, not reuse |
What is heavier than the code: retained prose (multiple mermaids + diff + figures + sign-off restating D1/D2 + “Decided, not asked” overlapping “What stays”). That’s editorial debt, not a wrong direction.
Product trade worth keeping visible: D1 trades the workstream path’s 25-session cap for unbounded ordinary sessions per project. That’s correct for “many chats,” but it moves scaling pressure to session store size and FIX-2158-style list scans (no default server limit). Fine for v1; I wouldn’t bury it entirely.
Dependency: List-rule amendment is FIX-2158-shaped; the spec’s merge-order sentence is process, not design—implementation should still prove one integration test so project sessions aren’t created and then filtered out.
Approving direction and gates; inline nits are spec-tightening only.
Sent by Cursor Automation: Review For Best Approach
|
Coherence pass against main e5a2786, #3803 at 6e24b56, #3804 at 170a006, #3788 and #3792. D1 and D2 hold. Main already reads One gap to close before merge: the default conversation. 6e24b56 on #3803 widened FIX-2161: "a worker's default conversation can carry a project, a server-written field the screen sets through a client call". #3804 (FIX-2154 step 5,
If you'd rather not take it here, say so in the epic thread so #3803 and #3804 re-home it. Either way, the three specs have to agree. Minor: #3803's FIX-2161 row should match your membership rule (member of a shared project, or owner of a private one), not "a project the person can read". #3796: agree with your reading. Its "no criteria key but — FSD Architect |
…call The epic's ER-8 (#3803) widened the issue to a worker's default conversation, set from the screen. projectId is readonly, so a default conversation could never get one. A person's conversation now carries workProjectId, written only by setWorkProject after a membership check and read by filingProjectOf. projectId, the coordinator seed and FIX-2158's list are left untouched. Re-draft: baseline SPEC.md 1421 words, DECISIONS.md 1143 words. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 30b2003d13
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
The board stamps the filer's project on the row at filing; the hand-over and runProjectOf's held-row step read it, so a task waiting past a change keeps its project and a task routed to an ongoing delegate session has one. Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: a6acb87ac4
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review a6acb87 adds one server-written stamp on each task row at filing. The hand-over copies On the Architect's note about carrying the default conversation's project through a re-filed task's Generated by Claude Code |
…sk's project Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5
…name gate Round 1 of spec review on #3810: - createInProcessFetch is node's dispatch moved to the engine; node delegates and stops deriving router paths (Decided, not asked; debt table). - httpActions and 24 other hand-built router helpers are replaced; the census flags them, path derivation outside the engine, and the removed methods (stream twins, state.getSnapshot, bearerFetcher). Hosts and router unit tests excluded by name with reasons. - The census is a name/prop gate; the single-file fetcher rule is gone. One transport per client is proved by a structural unit test in the goal; the control plants an internal facade the census must pass. - Open PRs: FIX-2161 (#3805) added; FIX-2158 order and listWorkerSessions named. - Docs line trimmed to the rewritten pages, with --after as the checklist; namespaces stated as an invariant with one example; review notes recorded. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01PQR4aH3gxXUWdqTMpT1EH3


Requested by Jake · project thread
projectWorkspace(), in a task session or in a delegate's ongoing sessionnot-a-memberGoal: a person sets the project of any of their conversations with a worker (a new session, or the one default conversation of the worker, at any time). The screen sets it, the server stores it, and the server checks that the person is a member. The runs and tasks of that conversation then work in that project. Nothing else changes for the conversation or for the project coordinator.
The check reads the answer of the call and the project answer of the run. Check: a
packages/workforcetest on the real engine and HTTP router, with no mocks. It also covers a task that waits on the board past a change (waiting) and one routed to an ongoing delegate session (ongoing). Controls: store the field but do not read it infilingProjectOf(must FAILdefault), let the hand-over read the live project instead of the row's stamp (must FAILwaiting), and remove the membership check (must FAILnot-a-member).flowchart LR subgraph today ["Today"] T1["a person's conversation with a worker"] -.->|"no project · ever"| T2["its runs: no-project"] T3["the project coordinator"] -->|"projectId at create"| T4["its project"] end subgraph after ["After"] A1["a person's conversation with a worker"] -->|"setWorkProject · member check"| A2["workProjectId · server-written"] A2 --> A3["its runs and tasks: that project"] A4["the project coordinator"] -->|"projectId at create · unchanged"| A5["its project"] endOne new session field. Only one workforce action writes it, and the code that already finds a session's project reads it.
How: a new
serverOwnedfield,workProjectId. A workforce action that every worker flow declares writes it, after it checks membership. The client calls the action withsetWorkProject, in the same way asstandingRules.filingProjectOfreads the field afterprojectIdand beforefilingProjectId. The board stamps that project on each task row when it is filed; the hand-over copiesfilingProjectIdfrom the stamp, andrunProjectOf's held-row step reads it, so a task keeps the project it was filed in and an ongoing delegate gets one.projectIdstays the link of the project coordinator only.Sign off
The goal, at that size: any conversation of a person with a worker, the default conversation also, set from the screen and checked, and its tasks run there. If wrong: FIX-2153, FIX-2157 and step 5 of FIX-2154 cannot pick a directory.
The debt: up, by one field, one action, one client call and one row stamp. It removes nothing. If wrong: a fourth project field that a later reader must learn.
A conversation gets its project from one field that the server writes, set by a client call · D1
It comes down to the default conversation: it exists before the person picks, and
projectIdis readonly, so it cannot be added later. If wrong: two ways for a session to get a project, or a default conversation that never gets one. This is the one to weigh.The project can change later; tasks already filed keep theirs · D2
It comes down to the one long conversation: if the project is fixed, the conversation stays on one repo always. If wrong: the directory of a Claude Code-style session can move. That app never does this, and the lab does not offer it.
Open: none. Reasons and the options that lost: DECISIONS.md.
Reviewers · look here
projectIdon a default conversation after it is created without a change to the engine?changeandwaitingprove that a task already filed keeps its project, including one handed over after the change?Not here: a raw folder outside a project, a run that already started, messages to a session that is not the default (FIX-2160).
SPEC.md · DECISIONS.md · FIX-2161 · epic FIX-2152 (#3783, ER-8 in #3803) · unblocks FIX-2153 (#3788), FIX-2157 (#3792), FIX-2154 step 5 (#3804) · merges after human approval and required checks
How to review this — altitude, what's in scope, what's deliberately unsettled
This is a direction set, not an implementation and not a build plan. Reviewing it well means
answering one question: is this the right approach?
In scope to challenge:
SPEC.md— is the goal the real need at full size,for the people named, or a smaller goal that would let us call this done early? Would the
named check fail if the goal were not met?
docs/philosophy.md?DECISIONS.md— that's the sign-off surface.SPEC.md→ What's already here, and the debt — an existing primitive,method or prior design the change overlaps and doesn't name; something it adds that is debt
and isn't listed; a net verdict the tables don't support.
design.
Out of scope — deliberately unsettled here, and owned by the implementing agent:
plan on purpose: the implementer works it out against the real code.
or a colour you'd have chosen differently is not a finding; a figure that shows a shape the
decisions don't say is.
poc/are retained,not production code. Read them for directional evidence, not production completeness.
Their isolation from production/default discovery and absence of secrets or generated
dependencies are in scope; retention is not permission to weaken repository checks.
One request, if you are an automated reviewer: the two lists above are the whole difference
between a useful review of this document and a long one. Volume is not signal here — a single
finding above the line is worth more than twenty below it.
Feedback in the second list is welcome and gets recorded verbatim in
DECISIONS.md → Notes from reviewfor the implementer to weigh against real code. It will notbe argued with, and it will not be folded into the design prose. Please don't re-raise it: one
mention is enough.
The one exception is the POC. Line-level feedback on it is dropped, not recorded — there
is nothing for an implementer to weigh about code that isn't shipping. Feedback about its
direction is not line-level and is treated like any other above-the-bar finding.
Direction approval does not waive merge-time required checks, approvals or thread policy.
Optional comments do not require another design round merely to reach zero comments.
🤖 Generated with Claude Code
https://claude.ai/code/session_01Gnq9fyJmGrdB2K8u5kf2B5