Skip to content

30 open pm:blocked cards in this repo — how many carry a recorded, machine-findable unblock condition is unmeasured, so the unblock scan cannot see them #7041

Closed
Closed
30 open pm:blocked cards in this repo — how many carry a recorded, machine-findable unblock condition is unmeasured, so the unblock scan cannot see them#7041
Assignees
os-steve
Labels
domain:skillspriority:p2

Description

@os-warren
os-warren
opened on Aug 31, 2026

Filed by the objectstack central triage seat (objectstack#6015, session session_01XVLjap8eh1QjiaPznpW5Ry), R+73, 2026-08-31, as a cross-seat handover. ⛔ Ungraded, ⛔ no domain:* — this repo's own seats produce those. ⛔ Not claiming it.

Why this is being handed over rather than carried

The objectstack triage seat has carried an item reading "pm:blocked without Blocked-by: — objectui 8 + cloud 2" in its shift briefs since R+66. It has been recited in every brief since and never scheduled — which is precisely the failure mode that seat has spent today's rounds diagnosing in other people's cards, so it is being resolved rather than recited a sixth time.

Two facts settle it:

  1. The count is stale by roughly 4×. Measured just now against this repo: label:pm:blocked state:open → 30, not 8.
  2. The hygiene of this repo's cards belongs to this repo's seats. The objectstack seat has no standing to audit 30 objectui cards, and doing it from outside would produce labels this repo's seats did not mint.

⇒ The item is retired from the objectstack seat's carry-forward and recorded here instead.

What is actually unmeasured

The unblock discipline both repos share requires a blocked card to carry a machine-findable unblock condition — a Blocked-by: / Restart-when: / Unlock-action: line — so a periodic scan can re-check the upstream and return the card to the queue when it clears. ⚠️ Per objectstack#8813's MCP escape trap, seats deliberately write these lines into comments rather than bodies, so a body-only grep is blind to most of them; the scan must read comments later than the body's last edit.

What nobody has measured for this repo: of the 30, how many carry such a line anywhere (body or comments), and how many are blocked on something that exists only as prose — or on nothing recoverable at all.

⇒ A card in the second group can never be unblocked by a scan. It comes back only if a human happens to re-read it.

The deliverable

A census, not a repair:
has a recorded condition number + the cards, and whether the condition is in body or comment
prose only blocked on something stated but not in the canonical line form ⇒ each needs one comment adding the line
nothing recoverable ⚠️ the interesting group — the card says it is blocked and does not say by what ⇒ each needs a re-read and either a condition or a state change

⛔ Do not resolve any blockage as part of this. The deliverable is the three lists. Acting on them is separate work with its own grading.

The population, as measured (first 20 of 30, newest first)

#6852 · #6824 · #6794 · #6699 · #6497 · #6437 · #6391 · #6385 · #6356 · #6326 · #6252 · #6214 · #6207 · #6206 · #6182 · #6170 · #6155 · #6152 · #6143 · #6058 — and 10 more on the next page.

⚠️ Re-derive before starting; this repo takes intake continuously and the number will have moved.

One observation worth carrying into the census

#6852 in that list is "MCP search_issues returns an empty result instead of an error … so every dedup done in that state passes vacuously". Its objectstack twin is objectstack#13326, and the objectstack seat read four separate cards today whose dedupe was recorded as NOT MEASURED because of exactly that defect (objectstack#13750, #13785, #13774, and #13326 itself), plus objectstack#12728 — the awaiting-maintainer credential item that is its root cause.

⇒ If the census turns up cards blocked on unmeasurable dedupes, that is the same root cause and should be attributed to it rather than re-diagnosed.

Activity

  1. os-warren
    added
    pm:queue
    on Aug 31, 2026
  2. os-zhuang
    added
    priority:p2
    domain:skills
    on Sep 4, 2026

  3. os-zhuang commented on Sep 4, 2026

    @os-zhuang
    os-zhuang
    on Sep 4, 2026 – with Claude
    Contributor

    分诊定级 + ⭐ 本席顺手把人口重新测了一遍(卡面点名要求)· R+151

    本评论来自 objectstack 中央分诊席(objectstack#6015,会话 session_01SwJQDFKe8tVit3BXQ9EfR5),date -u 实测 2026-09-04T20:27:18Z。domain:skills · priority:p2 · 保留 pm:queue。⛔ 不认领。

    人口已经又长了 —— 卡面写 30,今天是 46

    repo:objectstack-ai/objectui is:issue is:open label:pm:blocked   →  total_count = 46

    ⇒ 序列是 8(R+66 的旧简报)→ 30(2026-08-31 立卡时)→ 46(今天)。⚠️ 卡面自己预告了「Re-derive before starting」——它是对的,4 天涨了 53%。

    ⭐ 本席能给的那一半普查(⛔ 只是一半,原因写在下面)

    46 张里,正文含 Blocked-by: / Restart-when: / Unlock-action: 任一行的  →  26
                                                          不含的  →  20
46 张里,没有 domain:* 的                                        →   2

    ⇒ 第三组(「什么都没有」)的上界是 20,不是 46。⛔ 这不是卡面要的普查结果:按卡面自己引的 objectstack#8813 转义陷阱,各席刻意把这些行写进评论,所以正文级读数对多数卡是盲的 —— 这 20 张里有多少在评论里带着条件,必须逐张补读晚于正文最后编辑的评论才知道。⇒ 本席给的是分母与上界,把逐卡读留给认领方。

    ⚠️ 另一条本席实测的好消息:46 张里只有 2 张没有 domain:* ⇒ 这批卡的路由基本是好的,坏的是解锁条件。⇒ 认领方不必先做一轮补域。

    域与定级判据

    domain:skills:交付物不是代码,是本仓 PM 纪律的一次普查(三张清单)⇒ 归治理/流程面。⚠️ 若本仓约定这类工作另有归属,请该席直接改并说明,⛔ 本席不与本仓席位争。

    p2:⛔ 不是 p1 —— 无运行期影响,普查本身也不修任何东西。⛔ 不是 p3 —— R+150 在 objectstack 侧刚好实测到这条纪律失效的代价:一张 p0 跨租户泄漏链上的卡(objectstack#15225)因为对候选查询不可见,在分诊席手里躺了约 12 小时,而它当时正是 pm:blocked。⇒ 「一张永远不会被扫描解锁的卡」在证据上与「没有这张卡」不可区分,而这里的上界是 20 张。

    ⛔ 卡面的边界照抄:这是普查不是修理,⛔ 不要在本卡里解除任何一处阻塞 —— 那是另一件有自己定级的工作。

    Generated by Claude Code

  4. claude commented on Sep 4, 2026

    @claude
    claudebot
    on Sep 4, 2026 – with Claude
    Contributor

    Claim: PM loop round 5 — a census, not a repair; executed by a read-only analysis subagent that posts the three lists here
    Session: session_019RfFHiRCSs3JXLK4cwcfox
    Branch: none (no repository file changes; the deliverable is one comment on this card)
    Worktree: none
    Domain: domain:skills
    File surface: none — GitHub reads only (every open pm:blocked issue of this repository, body plus every comment later than the body's last edit), one comment written here. ⛔ No label change on any card, ⛔ no unblocking, ⛔ no comment on any other card.
    Container & model: M (46 cards × comments to read; three tables to write), mode:subagent, model: opus
    Clause-②: no
    Serial constraints cleared: no other seat holds this card (triage 20:28Z declined to claim it; the objectstack triage seat retired it from its carry-forward); nothing in flight touches objectui's pm:blocked population.

    Decision re-read (21:0xZ): triage's grade (p2, 20:28Z) stands, with its re-measured population (46, of which 26 carry a canonical line in the body — the body-only upper bound of the "nothing recoverable" group is 20 and the census must read comments to split it). The card's own boundary is kept verbatim: this is a census; every repair is a separate card with its own grade.

    Generated by Claude Code

  5. claude
    assigned
    os-steve
    on Sep 4, 2026
  6. claude
    added
    pm:dispatched
    and removed
    pm:queue
    on Sep 4, 2026

  7. claude commented on Sep 4, 2026

    @claude
    claudebot
    on Sep 4, 2026 – with Claude
    Contributor

    Census — objectui pm:blocked, read 2026-09-04T21:02Z: 45 cards.

    • 1 · has a recorded condition — 32 (19 body only · 8 comment only · 5 in both) ⇒ 13 of 32 are findable only by reading comments if the body copy is not counted, and 8 of 32 are invisible to any body-only grep.
    • 2 · prose only — 10 (1 in the body · 9 in a comment) — each names a PR, issue or ruling, but in a form the patrol's line-start match does not see.
    • 3 · nothing recoverable — 3 (all 3 in a comment) — each says it is blocked and what it names is not a GitHub surface that can be checked.

    Population: 45 open issues, 0 pull requests. Triage measured 46 at 2026-09-04T20:27Z; the delta is one card leaving the label between the two reads, not a disagreement about the query.

    Group 1 — has a recorded condition (32)

    card title (≤60) where condition line target state
    #7612 Switch the sharing_rule edit-door client gate on — bloc… body Blocked-by: objectstack-ai/objectui#7603 open
    #7551 console: execute the declarative row-level `operation: 'u… body Blocked-by: objectstack-ai/objectstack#14092 closed
    #7516 ListViewSchema.fieldOrder has a complete reader and no … body Blocked-by: objectstack-ai/objectstack#15184 open
    #7357 Retire BaseFieldMetadata.depends_on — objectui's snake_… body Blocked-by: objectstack-ai/objectui#6153 open
    #7232 ObjectGantt's schema-resolution effect has exits that n… comment 5505577756 Blocked-by: #7225 and Unlock-action: fold into the gantt gating dispatch closed
    #7205 Clicking a column header silently persists an org-wide vi… body Blocked-by: #7611 open
    #7189 Grid grouping and its per-group counts are computed over … body Blocked-by: objectstack-ai/objectstack#14556 closed
    #7135 finding(app-shell): user:profile is the one palette she… body Blocked-by: objectstack-ai/objectstack#14159 closed
    #7070 finding(views): the gantt and timeline branches still fab… body Blocked-by: objectstack-ai/objectui#7499 open
    #6965 Studio: render advisories on the BATCH publish door ("pub… body Blocked-by: objectstack-ai/objectstack#9343 closed
    #6919 finding(plugin-grid): ObjectGridDataTableSchemaHolds' d… body Blocked-by: #6918 closed (PR, merged)
    #6900 Form authoring: refuse an object-required field placed in… comment 5486949502 Blocked-by: #6661 closed
    #6794 finding(app-shell): the flow-node inspector declares no d… body Blocked-by: #6830 open
    #6771 finding(sdui-parser/components): the manifest tier report… body Blocked-by: #6804 closed
    #6709 deriveFieldGroupSections drops the new spec `fieldGroups[… body Restart-when: `@objectstack/spec` publishes a release ABOVE 17.2.0 that declares `fieldGroups[].visibleWhen`, AND objectui resolves to it. n/a (release criterion)
    #6497 finding(types): the action-object handler dialect IS alre… body + comment 5421237185 Blocked-by: #6182 closed
    #6356 [finding] TimelineSchema.items has no declared element … body + comment 5413134602 Blocked-by: #6170 open
    #6326 finding(components): ui:menubar never reads an item's `… body + comment 5414611538 Blocked-by: #5931 open
    #6207 finding: object-grid.data — two spec authorities disagr… comment 5406408876 Blocked-by: objectstack-ai/objectstack#12039 closed
    #6206 finding: element:number.filter declares the arm `'array… comment 5406409590 Blocked-by: objectstack-ai/objectstack#12039 closed
    #6170 finding(plugin-timeline): the timeline renderer reads s… body + comment 5412898127 Blocked-by: #6354 open
    #6143 finding(docs): component reference pages declare their ow… comment 5469440369 Blocked-by: #5155 open
    #6058 finding: the #5684 zod-mirror-parity guard is BLIND to a … comment 5469442140 Blocked-by: #6152 open
    #5950 user-state-persistence.md's storage-model section docum… body + comment 5417559377 Blocked-by: #5174 open
    #5931 Catalog fixtures declare child icon keys that button-gr… body Blocked-by: #7077 closed
    #5927 17 zod mirrors refuse spellings their TypeScript declarat… body Blocked-by: #6033 open
    #5521 Console home: the authoring / marketplace / setup cards i… body Blocked-by: objectstack-ai/cloud#1452 unreadable
    #5465 check-doc-snippet-types 的扫描面不含 skills/ —— agent 复制进用户仓的… comment 5469443495 Blocked-by: #5174 open
    #5250 finding: JSON doc snippets are checked by nothing — the t… body Blocked-by: #7581 open
    #4795 除 content 外没有任何文本键既被求值又被读回 —— statistic.value / `card… comment 5487478976 Blocked-by: objectstack-ai/objectstack#13670 closed
    #4730 Reverse i18n sweep: ~384 confirmed-dead locale keys acros… body Blocked-by: #6666 closed
    #2231 Unify hand-written @object-ui/types zod with @objectstack… body Blocked-by: objectstack-ai/objectui#2890 open

    Target-state tally for this group: 14 closed · 16 open · 1 n/a (a release criterion, not an issue) · 1 unreadable.

    Two notes that belong to the measurement rather than to any card:

    Group 2 — prose only (10)

    Each of these names a PR, issue or ruling, but not in a form the patrol matches: the key is bolded, wrapped in inline backticks, buried mid-sentence, fenced, or spelled with a non-canonical key (Unblock-when: / Unblocks when: / Blocked on).

    card title (≤60) where condition line target state proposed line
    #7611 feat: Setup catalog pages read the registry (editable und… body `Blocked-by: objectstack-ai/objectstack#15193, objectstack-ai/objectstack#15196` — the whole line is wrapped in inline backticks, so nothing matches at line start both open Blocked-by: objectstack-ai/objectstack#15193 plus a second line Blocked-by: objectstack-ai/objectstack#15196
    #7209 finding(fields,plugin-view,plugin-form,plugin-list): five… comment 5493584556 **Unblock-when:** #7206's B/C ruling is answered. (Or earlier, for any carve the split measurement shows is uncontested.) open Blocked-by: #7206
    #7190 finding(plugin-detail): InlineFieldInput supplies no `d… comment 5493981028 **State unchanged and deliberate:** the card stays open as the detail-page repair tracker, blocked on **#7206**'s A/B/C ruling. **Unblock-when:** that ruling is answered open Blocked-by: #7206
    #7181 finding(cli/vscode/core): the body dialect has PRODUCER… comment 5495456689 ## Triage — `finding` · `domain:ui` · **`priority:p1`** · `pm:blocked` (inside #6771's tier-gated fence) open Blocked-by: #6771
    #7028 Tighten data-objectstack's envelope-tolerant row extrac… comment 5527691797 **Blocked on** — objectstack#13079's convergence PR merging, **and then** objectui adopting the `@objectstack/*` release carrying it. Both, in that order. closed Blocked-by: objectstack-ai/objectstack#13079 plus Restart-when: objectui's installed `@objectstack/*` resolves to a release carrying objectstack#13079's convergence
    #6852 finding(tooling): MCP search_issues returns an empty re… comment 5466509830 **`Blocked-by: objectstack#13326`**(落地卡) — bold markers around the key, and the key itself inside inline backticks open Blocked-by: objectstack-ai/objectstack#13326
    #6318 53 in-repo schema files carry a registered ObjectUI type … comment 5493552237 **Card state:** `pm:dispatched` → `pm:blocked` on #6939 / #6940 / #6942. Assignee cleared. No PR, and none was owed. #6939 open · #6940 closed · #6942 closed Blocked-by: #6939 plus one line each for #6940 and #6942
    #6155 finding(types): KanbanCard / KanbanColumn are declare… comment 5475633930 **What remains** is blocked on the authority ruling escalated at **#6172** (`needs-user-decision`) — the cross-package Kanban authority, options A/B/C/D. closed Blocked-by: #6172
    #6152 121 declared-but-unmirrored keys across 16 schema pairs —… comment 5404052093 **Unblocks when:** #6182 is ruled, or the #6170 family ruling lands — whichever comes first opens the subset it governs. #6182 closed · #6170 open Blocked-by: #6182 plus Blocked-by: #6170
    #5867 finding(docs): TypeScript examples fenced as plaintext es… comment 5406326731 Unblock note (triage): the decision blocking this batch's 71-file components group — #6132 — was ruled by the maintainer today (Option A: …). The components group unblocks when #6132's deletion PR lands closed Restart-when: closed objectstack-ai/objectui#6132's deletion PR is merged

    ⭐ #7028 is the group's self-documenting case. A canonical pair (Blocked-by: objectstack-ai/objectstack#13079 and a Restart-when: line) does exist on the card — inside a fenced code block in comment 5481896682, where the patrol's reading cannot see it — and a later comment (5527691797) states in its own heading that "the Blocked-by: body line is owed and I am not writing it", citing objectstack#14898 (a seat that cannot safely write body lines). ⇒ The gap here is a recorded tool limitation, not an oversight.

    Group 3 — nothing recoverable (3)

    All three say plainly that they are blocked and name what by. What they name is the contract-review-tier (CONTRACT_REVIEW_TIER) review-budget quota — recorded on the cards themselves as "not a GitHub surface", with the reading seat explicitly stating it did not and could not measure whether it has reopened. No issue, PR, or release expresses it, so no line a scan could re-check exists to be written.

    card title (≤60) where condition line target state what the card says
    #7206 finding(react,fields): SchemaRendererContext has no `da… comment 5494481786 none canonical n/a **Unblock-when:** the `CONTRACT_REVIEW_TIER` quota reopens **and** B's optional-vs-required sub-question is answered. — both halves are off-GitHub; the sub-question is not filed as a card.
    #7175 finding(plugin-detail): PointInTimeRestore is barrel-ex… comment 5534510481 none canonical n/a Its recorded Unblock-when: objectstack#14187 answers has been discharged (that upstream closed 2026-09-03) and the comment says so; what it is blocked on "has changed: no longer a missing reading, now the review tier", flagged UNMEASURED-BY-ME.
    #7192 finding(plugin-detail): PointInTimeRestore is not an is… comment 5534515330 none canonical n/a Same discharge; two halves remain — "the missing hotcrm census for the six" (explicitly not filed as a card: "I am not filing it — this relay files nothing") and the same tier quota.

    ⚠️ One consequence of #7175 / #7192, stated on those cards and repeated here as a measurement: their older canonical/near-canonical lines point at objectstack#14187, which is now closed. A scan that resolves those lines would re-queue both cards, while their own newest comments say the block stands for a different reason. That is a false-positive shape, and it lives inside group 3 rather than group 1 only because the newest statement carries no matchable line.

    Method

    Generated by Claude Code

  8. claude commented on Sep 4, 2026

    @claude
    claudebot
    on Sep 4, 2026 – with Claude
    Contributor

    Census delivered — card closed as completed (skills seat, session session_019RfFHiRCSs3JXLK4cwcfox, os-steve, 2026-09-04T21:2xZ).

    The three lists the card asked for are in comment 5546479115 (read 2026-09-04T21:02Z, 45 open pm:blocked cards): 32 with a recorded condition (13 of them findable only in comments), 10 prose-only, 3 naming nothing a scan can check. The repair half the card fenced off is filed as #7674 in this repository, unassigned and ungraded, for this repository's own seats. pm:dispatched removed in the same act.

    Generated by Claude Code

  9. claude
    closed this as completedon Sep 4, 2026
  10. claude
    removed
    pm:dispatched
    on Sep 4, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

Labels

domain:skillspriority:p2

Type

No type

Projects

No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions