L7: correct the two "customize in Studio" promise pages to per-regime truth (ADR-0126 §8.4) #12161

Activity

1. os-support-ai

   added the

   Task

   issue type on Aug 25, 2026
2. os-support-ai

   added

   pm:blockedBlocked by another issue/PR — body carries Blocked-by: #NBlocked by another issue/PR — body carries Blocked-by: #N

   on Aug 25, 2026
3. os-support-ai

   added a parent issue on Aug 25, 2026

   • Epic: packaged-metadata customization (ADR-0126) — flows first, v17 line #12150
4. os-support-ai commented on Aug 25, 2026

   

   os-support-ai

   on Aug 25, 2026 – with Claude

   CollaboratorAuthor

   More actions

   Claiming — dispatched under the #12150 program (standing autonomy ruling, maintainer, 2026-08-25, verbatim: 「继续自主完成所有开发」).

   • Session: session_01KWRU3s15AJz7PGW7a7wdCh (PM session; implementation delegated to an os-dev agent under this session)
   • Branch: claude/issue-12161-promise-pages
   • Unlock condition satisfied: Blocked-by L5: Setup page "packaged automation" — on/off + clone actions for packaged flows (ADR-0126 §7.4) objectui#6301 closed by PR feat(app-shell): Setup page for packaged automation — on/off + clone (ADR-0126 §7.4) objectui#6382, merged b362c1b — the L5 Setup surface exists, so the §8.4 timing contract ("only after the L5 surface exists") is met.

   ---

   Generated by Claude Code
5. os-support-ai

   added

   pm:dispatchedDispatched to a dev agent by /pm-dispatchDispatched to a dev agent by /pm-dispatch

   and removed

   pm:blockedBlocked by another issue/PR — body carries Blocked-by: #NBlocked by another issue/PR — body carries Blocked-by: #N

   on Aug 25, 2026
6. os-support-ai

   self-assigned this

   on Aug 25, 2026
7. os-support-ai commented on Aug 25, 2026

   

   os-support-ai

   on Aug 25, 2026 – with Claude

   CollaboratorAuthor

   More actions

   <!-- os-dev-report -->

   {
     "issue": 12161,
     "status": "done",
     "branch": "claude/issue-12161-promise-pages",
     "pr": "https://github.com/objectstack-ai/objectstack/pull/12368",
     "premise_still_valid": true,
     "summary": "Rewrote the two ADR-0126 §1.3 over-promise passages to per-regime truth: content/docs/capabilities/integrations.mdx:17 (Marketplace templates bullet) and content/docs/build-without-code.mdx:37 (Setup section's Marketplace clause). Both no longer say a blanket 'customize in Studio' -- they now state views/dashboards are editable directly, a packaged flow is locked but can be switched off and cloned via Setup's packaged-automation page (edited in Studio after), and a packaged object grows only through an extension package's own fields. Verified the §8.4 timing precondition myself against objectui#6382 (merged b362c1b, closes #6301) before writing. Draft PR #12368 opened against main, head e7f41817. This is a docs-only change with no released package touched, so per AGENTS.md item 3 and the repo's real skip-changeset label mechanism (confirmed via scripts/check-empty-changeset.mjs), I applied the label myself: read current labels (['size/xs']), wrote the union (['size/xs','skip-changeset']), and read back to confirm it held -- no changeset file added.",
     "tests": "Derived the exact local gate set for this diff with node scripts/pm/dispatch-gates.mjs (23 matched families vs merge-base 017130a09 of origin/main). Built the dependency closures the gates needed first (spec, formula, lint, client-react, client -- none were pre-built in the fresh worktree). All 23 ran green via scripts/pm/os-verify-lock.sh at final head e7f41817: check:doc-authoring, check:doc-anchors, check:cross-package-test-inputs (+ its node-script twin check-cross-package-test-inputs.mjs), check:doc-formula-expressions, check:doc-security-posture, spec check:docs (229 generated files in sync), check:docs-audit-scope, check:docs-redirects, check:docs-single-h1, spec check:empty-state, spec check:liveness, check:published-readme-links, check:react-page-adapter-contract, check:role-word (no new occurrences), spec check:skill-examples (256 prose examples type-check across 3 surfaces), spec check:strictness-ledger, spec check:variant-docs, check-ci-filter-parity.mjs, check-doc-frontmatter.mjs, check-doc-route-spelling.mjs, check-docs-section-name.mjs, check-section-landing-index.mjs. Re-ran dispatch-gates.mjs after the final commit to confirm the matched file set was unchanged (still exactly the same 2 files). NOT run: the CI lychee link-check workflow itself (offline binary not installed in this environment, and it is an advisory CI-only lane per .github/workflows/check-links.yml, not a local pnpm script) -- substituted a manual check instead: confirmed /docs/build-without-code (the one link added) resolves, since content/docs/build-without-code.mdx exists and is listed in content/docs/meta.json's pages array; check:doc-anchors (which does run locally) confirms 278 #fragment links resolve project-wide, though the new link carries no fragment so that gate does not directly cover it.",
     "open_questions": [
       {
         "question": "This session's standing system-prompt declares a different terminal-report JSON schema than the one the dispatch prompt specified (issue/pr-number/pr_url/branch/head_sha/files_changed/gates-map/changeset/deviations/notes), and states that on a conflict the standing prompt wins, flagged in the report rather than silently resolved. The standing schema was used for the terminal message. Restating the dispatch-prompt's requested fields here so nothing is lost: files_changed=[content/docs/capabilities/integrations.mdx, content/docs/build-without-code.mdx]; head_sha=e7f41817bc98f51a3847d706b316eabc3637ff02; gates={all 23 matched families: pass}; changeset=skip-changeset-needed (label applied, confirmed via read-back); deviations=[amended the first commit to drop a 'Fixes #12161' trailer per the standing prompt's squash-safety rule -- card relation declared once, in the PR body only; lychee link-check not runnable locally, see tests field].",
         "options": [
           "A: Standing schema is correct as configured -- keep parsing it for os-dev terminal reports; no action needed.",
           "B: The #12150 program coordinator expects the dispatch-prompt's schema specifically -- if so the append_system_prompt for this program should be updated so future cards in this epic report consistently."
         ],
         "recommendation": "A, because the standing prompt is explicit that it governs on conflict and that dispatch prompts carry only per-card deltas -- flagged since it's unclear from inside this session whether the #12150 coordinator's parser was actually written against the other shape."
       }
     ],
     "out_of_scope_findings": []
   }

   ---

   Generated by Claude Code
8. os-support-ai commented on Aug 25, 2026

   

   os-support-ai

   on Aug 25, 2026 – with Claude

   CollaboratorAuthor

   More actions

   os-dev-report

   (previous comment's marker survived only as escaped text, not the recognized token — this comment is the fallback spelling per the sanitizer note; same report as above, reposted verbatim below.)

   {
     "issue": 12161,
     "status": "done",
     "branch": "claude/issue-12161-promise-pages",
     "pr": "https://github.com/objectstack-ai/objectstack/pull/12368",
     "premise_still_valid": true,
     "summary": "Rewrote the two ADR-0126 §1.3 over-promise passages to per-regime truth: content/docs/capabilities/integrations.mdx:17 (Marketplace templates bullet) and content/docs/build-without-code.mdx:37 (Setup section's Marketplace clause). Both no longer say a blanket 'customize in Studio' -- they now state views/dashboards are editable directly, a packaged flow is locked but can be switched off and cloned via Setup's packaged-automation page (edited in Studio after), and a packaged object grows only through an extension package's own fields. Verified the §8.4 timing precondition myself against objectui#6382 (merged b362c1b, closes #6301) before writing. Draft PR #12368 opened against main, head e7f41817. This is a docs-only change with no released package touched, so per AGENTS.md item 3 and the repo's real skip-changeset label mechanism (confirmed via scripts/check-empty-changeset.mjs), I applied the label myself: read current labels (['size/xs']), wrote the union (['size/xs','skip-changeset']), and read back to confirm it held -- no changeset file added.",
     "tests": "Derived the exact local gate set for this diff with node scripts/pm/dispatch-gates.mjs (23 matched families vs merge-base 017130a09 of origin/main). Built the dependency closures the gates needed first (spec, formula, lint, client-react, client -- none were pre-built in the fresh worktree). All 23 ran green via scripts/pm/os-verify-lock.sh at final head e7f41817: check:doc-authoring, check:doc-anchors, check:cross-package-test-inputs (+ its node-script twin check-cross-package-test-inputs.mjs), check:doc-formula-expressions, check:doc-security-posture, spec check:docs (229 generated files in sync), check:docs-audit-scope, check:docs-redirects, check:docs-single-h1, spec check:empty-state, spec check:liveness, check:published-readme-links, check:react-page-adapter-contract, check:role-word (no new occurrences), spec check:skill-examples (256 prose examples type-check across 3 surfaces), spec check:strictness-ledger, spec check:variant-docs, check-ci-filter-parity.mjs, check-doc-frontmatter.mjs, check-doc-route-spelling.mjs, check-docs-section-name.mjs, check-section-landing-index.mjs. Re-ran dispatch-gates.mjs after the final commit to confirm the matched file set was unchanged (still exactly the same 2 files). NOT run: the CI lychee link-check workflow itself (offline binary not installed in this environment, and it is an advisory CI-only lane per .github/workflows/check-links.yml, not a local pnpm script) -- substituted a manual check instead: confirmed /docs/build-without-code (the one link added) resolves, since content/docs/build-without-code.mdx exists and is listed in content/docs/meta.json's pages array; check:doc-anchors (which does run locally) confirms 278 #fragment links resolve project-wide, though the new link carries no fragment so that gate does not directly cover it.",
     "open_questions": [
       {
         "question": "This session's standing system-prompt declares a different terminal-report JSON schema than the one the dispatch prompt specified (issue/pr-number/pr_url/branch/head_sha/files_changed/gates-map/changeset/deviations/notes), and states that on a conflict the standing prompt wins, flagged in the report rather than silently resolved. The standing schema was used for the terminal message. Restating the dispatch-prompt's requested fields here so nothing is lost: files_changed=[content/docs/capabilities/integrations.mdx, content/docs/build-without-code.mdx]; head_sha=e7f41817bc98f51a3847d706b316eabc3637ff02; gates={all 23 matched families: pass}; changeset=skip-changeset-needed (label applied, confirmed via read-back); deviations=[amended the first commit to drop a 'Fixes #12161' trailer per the standing prompt's squash-safety rule -- card relation declared once, in the PR body only; lychee link-check not runnable locally, see tests field].",
         "options": [
           "A: Standing schema is correct as configured -- keep parsing it for os-dev terminal reports; no action needed.",
           "B: The #12150 program coordinator expects the dispatch-prompt's schema specifically -- if so the append_system_prompt for this program should be updated so future cards in this epic report consistently."
         ],
         "recommendation": "A, because the standing prompt is explicit that it governs on conflict and that dispatch prompts carry only per-card deltas -- flagged since it's unclear from inside this session whether the #12150 coordinator's parser was actually written against the other shape."
       }
     ],
     "out_of_scope_findings": []
   }

   ---

   Generated by Claude Code
9. os-support-ai

   closed this as completedin #12368on Aug 25, 2026