Repository navigation
Description
Is there an existing issue for this?
- I searched existing issues for SDD/branch/circular clarification and no matching issue was returned. Size recovery is separately tracked in763.
Which actions are affected?
Issue
Which platforms are affected?
Linux
Description
A pre-branch SDD question asks the maintainer to confirm that the Action has already published the exact managed issue branch before the SDD preparation can continue. The same question card says no SDD draft or branch exists yet. This promotes an ordinary post-publication implementation gate into a circular pre-publication clarification, although branch creation/readiness is owned by the Action.
Expected: retain the requirement that source implementation waits for the exact admitted managed branch, but do not require that branch as a missing product decision before publishing its prerequisite SDD. Treat this as an orchestration postcondition, or ask whether the contributor will wait for it. Do not require a false confirmation, contributor-created ref or weakened source-change guard.
Reproducing the issue
- Consumer ✨🧑💻 - Complete household administration APIs and workspace over verified schema9 profile storage domus#25, implementation-only continuation after actual PR24/schema9 merge6d84d4815a8bbfe071979df6650d01e433dcc2db.
- Read question card ✨🧑💻 - Complete household administration APIs and workspace over verified schema9 profile storage domus#25 (comment) . It says Current status Waiting for specification answers, no SDD draft or branch exists yet, owning SDD specs/household-administration.md, clarification round1.
- Q1 validly asks whether PR24 merged and its tree matches reviewed head4e59e04d956668c76f3c8d48e824059fd205d661. This was answered from real Git evidence: both have treeca029dc8e1d90a566cd2031203af8992b7ba26eb.
- Q2 asks: Before source changes start, has the Action admitted this issue and published its exact managed branch? Suggested answer: Confirm the managed branch is available before implementation.
- Embedded SDD gate record has phase awaiting-answer, baseSha6d84d4815a8bbfe071979df6650d01e433dcc2db, action update, round1 and the two maintainer questions, but no published branch/commit. At that observed point no draft/branch was claimed published.
- The installed contributor contract says wait for pre-branch SDD publication and the Action-owned exact branch before source work. The consumer issue explicitly repeats that normal gate; it is not a request for the contributor to create a branch.
copilot Version
Consumer configured vypdev/copilot@v3. The adjacent original size attempt41462963263/job125342027777 uses the released pin recorded in the attached source diagnostics; the question card itself is the primary observation. No newer upstream fix, deterministic failure rate, hidden prompt contents or exact question-producer job is inferred.
Relevant Log Output
Current status: Waiting for specification answers. No SDD draft or branch exists yet.
Q2 Maintainer: Before source changes start, has the Action admitted this issue and published its exact managed branch?
Suggested answer: Confirm the managed branch is available before implementation.
phase=awaiting-answer; round=1; baseSha=6d84d4815a8bbfe071979df6650d01e433dcc2db
Additional context and comments
Maintainer recovery answers honestly: admission/start are verified, branch does not yet exist, the Action should publish the supported SDD and exact branch first, then source work can start after independent readiness/base verification. No branch is fabricated and no Copilot source/resource/configuration/workflow/credential/limit is modified. A separate terminal18,000-character size error is being recovered on this same issue through a complete new unregistered companion under the unchanged owner, checked with seven byte-verified original SDK modules. Local proposal validation is not native readiness. Healthy queued/running jobs are not restarted or canceled.
Suggested regression cases: a genuine unresolved product policy, missing actual merge evidence, already verified merge/tree, an admitted issue before branch creation, an existing verified managed branch, a post-publication branch failure, unrelated/missing contributor ref, and future implementation-only postconditions. Preserve human authorization and the no-source-work-before-readiness guard. This reports an observed circular question; it does not assert that truthful numbered answers can never recover the flow.
Metadata
Metadata
Assignees
Labels
Type
Projects
- StatusShow more project fieldsDone
Activity
[-]Pre-branch SDD clarification asks for its Action-owned branch to already exist[/-][+]🐛 - Pre branch SDD clarification asks for its Action owned branch to already exist[/+]vypbot commentedon Nov 2, 2026
Hi! I’m @vypbot, the Copilot assistant for this repository.
I can answer questions, explain the codebase, propose implementation and test plans, review issues and pull requests for potential bugs or security problems, and help authorized maintainers apply changes.
Try
/copilot helpto see the available commands, or mention @vypbot with your question.efraespada commentedon Nov 2, 2026
Recovery outcome: after the honest numbered answers vypdev/domus#25 (comment) and the supported new-companion proposal, the Action published171244421fcc6289180bdb94bf4e7bb52a990a1b with exact parent6d84d4815a8bbfe071979df6650d01e433dcc2db. Gate vypdev/domus#25 (comment) is phase=published with questions=[] and the exact feature/25-complete-household-administration-apis-and-workspace-over-verified-schema9-profile-storage branch. Branch, parent and the three documentation-only publication changes were verified before implementation. This proves supported recovery worked; it does not establish that the circular prerequisite question was valid or that the underlying defect is corrected. No unpublished branch was fabricated, no false prior-readiness answer was supplied, and Copilot remains unchanged.
A separate later terminal invocation41466725960/job125349518671 fails requirement-mapping/e34d3ce7-0abc-4e56-ba1b-4c776fc06b4b while that published gate/ref remain present, leaving blocked/awaiting-maintainer labels without branched. This later consistency/reconciliation observation is being retained with the existing mapping/source reports, not described as the same proven cause as the original circular question.