Blog: three second-wave ontology posts, and published posts that contradict the ontology guardrails (2 to correct, 4 lineage notes, 10 scenario sections) #236

Activity

1. objectstack-fleet

   added

   pm:queue

   content:refresh

   content:new

   pm:dispatched

   and removed

   pm:queue

   on Oct 15, 2026
2. objectstack-fleet

   assigned

   os-tesla

   on Oct 16, 2026
3. objectstack-fleet commented on Oct 16, 2026

   

   objectstack-fleetBot

   on Oct 16, 2026

   ContributorAuthor

   More actions

   Claim: PM loop round 1
   Session: session_017GeNDQH7E5S2rX8ajVAPxv
   Account: os-tesla
   Branch: claude/sharp-mendel-1087cg-issue-236-a
   Worktree: www.objectos.ai-issue-236-a
   Domain: none — single-lane repo (repo:www.objectos.ai)
   Seat: repo:www.objectos.ai#2
   File surface: item A only — new content/blog/ directory for post A ("A world, hands and a fence", slot Tue 2026-11-03), en + zh-Hans authored, zh-Hant generated (stop on breach; explain in the report). Items B, C and D stay on this card for later claims; the PR carries Refs #236 (item A), never Fixes.
   Container & model: M, mode:subagent, model: CONTRACT_REVIEW_TIER — reason: the maintainer's 2026-10-15 ruling that article cards dispatch at the ceiling tier, and #209's dispatch-tier ruling for positioning copy (2026-10-05)
   Clause-②: no
   Responsibility: n/a — not a defect card
   Thread-read: none
   Serial constraints cleared: slot 2026-11-03 is free (seat 1's plan on #209, comment 6192871332: 10-20 #244, 10-22 #245, 10-27 #237, 10-29 held for #231; recorded on objectstack-ai/objectstack#12224, comment 6209591407); new post directory, no file shared with any open PR; overlap of subject with #237 (agent-operable), #245 (AI-writable) and #244 (executable vs knowledge ontology) is fenced in the dispatch.
4. objectstack-fleet commented on Oct 16, 2026

   

   objectstack-fleetBot

   on Oct 16, 2026

   ContributorAuthor

   More actions

   os-dev-report

   {
     "issue": 236,
     "status": "done",
     "branch": "claude/sharp-mendel-1087cg-issue-236-a",
     "pr": "https://github.com/objectstack-ai/www.objectos.ai/pull/260",
     "session": "session_017GeNDQH7E5S2rX8ajVAPxv",
     "premise_still_valid": true,
     "summary": "Item A only. New post directory content/blog/what-do-ai-models-lack-in-the-enterprise/ with the post 'What Do AI Models Lack in the Enterprise? A World, Hands and a Fence' (en authored; zh-Hans re-derived; zh-Hant generated), date 2026-11-03, topic ai-agents, audience it, plus cover and diagram as zh / -en SVG pairs. The front page's line (apps/docs/app/[lang]/page.tsx:127-131) is mapped to the runtime mechanism supplying each of world, hands and fence with file:line at objectstack 1ecb08a1; one number measured with the README's own command on examples/app-crm (1,844 lines / 30 files; the fence declaration is 78 lines; 0 lines of tool/authorization/audit plumbing, the two grep hits are comments); the RAG + function calling + policy-layer stack is steelmanned with a five-row comparison; five limits are stated (REST /actions is outside the ai.exposed gate, read audit narrower than write audit, world = apps the runtime runs, action bodies are trusted code, no judgment supplied). Positioning lines quoted byte-exact from src/lib/positioning.ts (checked by script). Draft PR #260 opened through the fleet-write relay with assignee os-tesla and body 'Refs #236 (item A)'; items B, C, D untouched. PM mechanism assumption held, with the two bounds above written into the post.",
     "tests": "At HEAD 2bb6ff4 (final commit; the full gate set was rerun on it): pnpm content:lint exit 0 ('✓ content lint passed (350 files, 44 glossary terms checked)'; advisories only: title 68 cols, description 223 cols, 'Date is in the future' as expected). svg-text-fit alone on the four new SVGs exit 0 ('4 SVGs, 114 text boxes, 0 filtered elements, tolerance 2px, 0 allowlisted'). WCAG check script over all 24 text/background pairs: minimum 4.77:1, 'CONTRAST PASS'. Positioning byte-exact check: en T2/T4/E1 and zh t2/t4/e1 all 'ok'. Control-character scan (grep -naP) over the 7 files: none. Browser check (Playwright library at /opt/node-tools + Chromium /opt/pw-browsers/chromium against astro preview of dist): 'BROWSER CHECK PASS: 67 checks, 0 failing' — en/zh-Hans/zh-Hant at 1440x750 and 390x844, HTTP 200, img.cover=1 and 0 cover refs in body, all images scrolled into view and decode()d, 56/55/55 internal links all 200, no horizontal overflow, all tables overflow-x:auto scroll containers (mapping table revised to 3 columns so it fits at 390), blog list en + zh-Hans both widths show the post first with cover decoded; SVGs rendered at native size with headless Chromium and looked at, plus viewport/element crops at both widths looked at. No ablation: no tested code behaviour changed (content-only diff).",
     "gates": "Local subset, exits captured before any pipe, heavy ones under os-verify-lock.sh (OS_VERIFY_LOCK_SLOT=www-236a; first run waited 77s behind issue-235's holder, held 116s shared-box; second run lock free): build&&check at 2bb6ff4 → 'heavy-exit=0 HEAD=2bb6ff4', lock 'VERDICT batch-last-exit 0'. pnpm build exit 0: '✓ content lint passed (350 files…)' (--published), 'svg-text-fit: OK — … (122 SVGs, 2587 text boxes, 183 filtered elements, tolerance 2px, 1 allowlisted)', '[build] Complete!', '✓ share images: 91 rendered at 1200x630 … no uncovered glyph', '✓ content lint --dist passed (489 built blog pages checked)', '✓ content lint --dist CJK line breaks passed (891 built pages, …, 0 segment breaks…)', '✓ content lint --dist quotation marks passed (891 built pages, 12441 curly marks, 0 pointing the wrong way or left open…)', '✓ content lint --dist links passed (94461 internal links in 891 built pages resolve to built files)'. pnpm check exit 0: 'Result (154 files): 0 errors, 0 warnings, 0 hints'. pnpm seo:smoke exit 0: 'SEO smoke test passed (881 HTML pages checked; 580 sitemap URLs against 301 noindexed pages; 5508 canonical/hreflang annotations; 16524 prose blocks language-checked; 94451 internal links redirect-checked)'. No dispatch-gates.mjs derivation: that script lives in objectstack and this is www.objectos.ai, whose gate list was hand-derived from package.json as the dispatch prescribed. CI check-runs on PR #260 head 2bb6ff4 as read at report time: Build and deploy: success. Not waited for; the seat reads CI convergence.",
     "line_budget": "n/a",
     "files_changed": [
       "content/blog/what-do-ai-models-lack-in-the-enterprise/index.mdx",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/index.zh-Hans.mdx",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/index.zh-Hant.mdx (generated)",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/cover.svg",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/cover-en.svg",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/three-supplies.svg",
       "content/blog/what-do-ai-models-lack-in-the-enterprise/three-supplies-en.svg"
     ],
     "deviations": [
       "Sources were read at objectstack origin/main 1ecb08a1 (fetched today in the shared checkout; a fetch, no edit), not 5b6c257b named in the dispatch; the cited lines are identical at 712361b7 for page.tsx and the post names 1ecb08a1 throughout.",
       "Browser automation used the Playwright library already installed at /opt/node-tools (no playwright install, no npm install) with the preinstalled Chromium; one of my own helper scripts hung on img.decode() of lazy images it had not scrolled into view — the trap the dispatch names — so its first run was killed by PID (my processes only; a sibling's astro preview on port 4338 was left alone), the helper fixed to scroll first, and rerun. The first browser-check run also reported 9 false FAILs from a key typo in my checker (t.scrollWidth vs t.scrollW); the measured values already showed the tables scrolling; fixed and rerun to a clean PASS.",
       "Three fetch refs pr-244, pr-245, pr-237 were created as local branches in the shared www.objectos.ai .git to read the sibling posts; deleted at cleanup.",
       "The lock's first run sequenced build and check with ';' plus a final [ $B -eq 0 ] && [ $C -eq 0 ] (so the verdict did cover both); the second run on the final head joined them with '&&'.",
       "The mapping table went from 4 columns to 3 after the phone-width look (a 4-column table with long code spans rendered ~2,600px tall at 390); the fix was committed, and every gate rerun on the new head."
     ],
     "mcp_calls": "0 — no MCP tools called",
     "api_writes": "2 relay actions / 3 REST writes, all as objectstack-fleet[bot] through scripts/pm/fleet-write/dispatch.mjs: (1) pr_create = POST /repos/objectstack-ai/www.objectos.ai/pulls + its assignee leg POST /repos/objectstack-ai/www.objectos.ai/issues/260/assignees (run 38866809498, read-back identical 13091 bytes, assignees os-tesla); (2) comment = POST /repos/objectstack-ai/www.objectos.ai/issues/236/comments (this report). git push x4 to the branch (not REST). No label writes, no PATCH, no issue creation.",
     "open_questions": [],
     "out_of_scope_findings": [
       "none filed. Noted, not filed: content-lint's snippet-width advisory fires for any title over 60 columns (post 1's is 87; this one is 68, a query title per rule 2) — an advisory by design, not a defect. carrier: none."
     ]
   }

   ---

   Generated by Claude Code
5. objectstack-fleet commented on Oct 16, 2026

   

   objectstack-fleetBot

   on Oct 16, 2026

   ContributorAuthor

   More actions

   Split and status · 2026-10-16T07:37Z

   • Item A — PR content(blog): What Do AI Models Lack in the Enterprise? A World, Hands and a Fence (slot 2026-11-03) #260, seat ACCEPT (PR comment 6210408164), held in draft until its slot, Tue 2026-11-03.
   • Item B → Post (slot Thu 2026-11-05): When Should You Not Use an Executable Business Ontology? #261 (slot Thu 2026-11-05), dispatched.
   • Item C → Post (slot Tue 2026-11-10): Views, Dashboards and Apps Are Projections of the Ontology — What That Changes for Review #262 (slot Tue 2026-11-10), dispatched.
   • Item D (published posts against the guardrails) stays on this card. It is serial after seat 1's wave PRs (content(blog): agent-operable apps — every object and action an MCP tool under the user's permissions (post 4, published 2026-10-26) #237 content(blog): executable ontology vs knowledge ontology — post 2 of the positioning wave (slot 2026-10-18) #244 content(blog): what makes a business app AI-writable — post 3 of the positioning wave (published 2026-10-23) #245) and claimed when they have landed.

   This card keeps pm:dispatched and the seat-2 assignee while item A's PR is open. Under the partial-landing rule, it goes back to pm:queue when #260 merges, unless D is claimed by then.

   ---

   Generated by Claude Code
6. objectstack-fleet commented on Oct 17, 2026

   

   objectstack-fleetBot

   on Oct 17, 2026

   ContributorAuthor

   More actions

   Cadence change: one post a day, one shared calendar · ruling B on #267

   The maintainer ruled on 2026-10-17 (#267, verbatim answer 「B」) that the site moves to one post a day. Daily publishing starts once at least 7 approved posts are queued across all seats. Until then, posts keep their dated slots, with no gap longer than 3 days.

   What this means for this card's posts:

   • Slot calendar. Every seat now uses the calendar comment on Publishing cadence: one post a day, switched on once seven approved posts are queued (maintainer ruling B, 2026-10-17) — charter change, shared calendar, writing capacity #267. Your three PRs (content(blog): What Do AI Models Lack in the Enterprise? A World, Hands and a Fence (slot 2026-11-03) #260 on 11-03, content(blog): When Should You Not Use an Executable Business Ontology? #264 on 11-05, content(blog): views, dashboards and apps are projections of the ontology — the naming test, measured on app-todo (slot 2026-11-10) #263 on 11-10) are entered there as dated.
   • Claiming slots. Please claim future slots there before writing a date, so no two posts land on the same day.
   • Switch-over. When the queue reaches 7, the queued posts are re-dated onto consecutive days, and back-linked updated values move with them. Seat 1 coordinates the switch-over through Publishing cadence: one post a day, switched on once seven approved posts are queued (maintainer ruling B, 2026-10-17) — charter change, shared calendar, writing capacity #267 and does not edit another seat's PR.

   The charter change to AGENTS.md is a governed PR, which the maintainer merges.

   ---

   Generated by Claude Code
7. objectstack-fleet commented on Oct 21, 2026

   

   objectstack-fleetBot

   on Oct 21, 2026

   ContributorAuthor

   More actions

   Publishing moves to seat 1 · ruling on #267 (2026-10-21)

   The maintainer ruled 「发文章都归你这个席位」: publishing every post belongs to the repo:www.objectos.ai seat 1. That covers this card's three PRs, #260, #264 and #263. The cadence is now three a week, Mon/Wed/Fri (「一周3篇 周一周三周五」), and no new posts are written for now (「暂时只发不写新的」).

   On each publish day, seat 1 will:

   • add one commit to the PR's branch moving its date to that day (history is never rewritten);
   • review the post with a six-viewport check in both themes;
   • mark the PR ready;
   • merge it.

   Day	PR
   Wed 11-04	#260
   Mon 11-09	#264
   Fri 11-13	#263

   Please don't merge or re-date these yourselves. The single calendar is on #267.

   ---

   Generated by Claude Code