@bongos/core 1.21.169 → 1.21.171

package/docs/modules-contract.md

@@ -394,6 +394,7 @@ Run `bongos upgrade` to confirm coreVersion compatibility, then `node scripts/gd
 - Gate: Metic+ (`module.submit`) for now. This is not `bongos module submit`, which proposes a module *into* core.
 - **A how-to is required** ([ADR 0347](adr/0347-every-store-module-ships-a-how-to.md)). `modules/<key>/HOWTO.md` is what a buyer reads — the store shows it as the module's page — so it is written for the person who installs the module, not for an AI working inside it (that is `CLAUDE.md`). It is plain Markdown: any AI or person can write it. Publish refuses the version unless it has these five level-2 headings, in any order and any case, each with some text under it: `## What it does`, `## Install and enable`, `## How to use it` (with one worked example), `## Configuration` ("None." is fine) and `## Limits and known issues` ("None known." is fine). HTML comments don't count as text, so the scaffold's prompts must be replaced. The refusal names each failing section. The check is one function, `checkHowto` in `scripts/gds/module-artifact.js`; the CLI runs it before uploading and the store runs it again on upload, so a hand-built upload can't skip it. `bongos module check` reports it as advice only — an upstream submit is not gated by it. Install and update don't re-check it, so a version published before the gate still installs. The file travels in the tarball, so its hash pins it to the version. An optional `howto.artifactUrl` in `module.json` (an `https://` link on `claude.ai`) is shown as an extra, never graded.
 - **Getting listed** (task 1003807, [ADR 0343](adr/0343-a-module-score-is-a-security-gate-then-an-average-of-visible-parts.md)). Publishing a version does not by itself put it in the store's catalog, `GET /store/modules`. The catalog shows each module's newest version, judged by its current score. It must pass the Security gate first; a version that failed it, or has not been checked yet, is not listed. A new version with no install or reliability data yet is listed with a **New** label and is not ranked; the floor does not apply to it until it has a real score. Every other version is listed only if its overall score is **at least 50 out of 100**. Below that it is not listed at all. The floor is fixed in Bongos itself (`QUALITY_FLOOR` in `src/bongos/module-listing.js`) and returned in the catalog response; no author or instance setting moves it, and changing it is an owner-approved, reviewed task.
+- **Tester feedback** (task 1003806, ADR 0344 D5). While a version is on the tester channel, what tester projects report about it becomes its Tester feedback part: the share of testers whose install worked and who hit no runtime errors, blended with ten neutral testers so a small sample cannot score high — `round(100 × (went well + 5) / (testers + 10))`. Reports from after the version left the tester channel, and the author's own, never count; under 3 testers it is not scored yet. The rule is `assessTesterFeedback` in `scripts/gds/module-assess-installs.js`.
 - **What a module is strong and weak at** (task 1003799, ADR 0343 D4). Each catalog entry also carries `parts` (the current result of Tests, Clean installs, Reliability, Tester feedback and How-to) and a reading of them: `strengths` (a part scored **80 or more**), `weaknesses` (a part scored **below the floor**) and `unmeasured` (a part with no score yet, which is never counted as weak). Each strength and weakness has a one-sentence `detail` saying what the number rests on. Security is not among them: every listed version has passed it. The reading is `readParts` in `src/bongos/module-listing.js`; the hall's Modules tab shows it in its "In the store" group, so two modules with the same overall score can be told apart.
 - **Sorting by a part** (task 1003800). `GET /store/modules?sort=<part>` orders the listed modules by one part a buyer cares about: `tests`, `install`, `reliability`, `tester_feedback` or `docs` (the response lists them as `sort_parts`, and says which one it applied as `sort`). Modules with a score for that part come first, highest first; a tie goes to the higher overall score, then the module key. A module with no score for that part goes last, so missing data never wins or loses a place on its own. New modules are not sorted: they are never ranked. Any other `sort` value is refused with a 400. Without `sort`, both lists are in key order. The rule is `sortByPart` in `src/bongos/module-listing.js`.
 - **Testing new versions first** (task 1003802, [ADR 0344](adr/0344-a-tester-is-a-project-that-opts-in-once.md) D1). A project can choose to get each new version of its modules before everyone else. It is the project that opts in, never one builder for someone else's project, and never by default. `GET /modules/testing` returns `{ tester, general_only, updated_at, updated_by }` to anyone signed in. `POST /modules/testing { tester: true|false }` switches it, and `POST /modules/:key/general-only { general_only: true|false }` keeps one module on the versions everyone gets while the rest are tested. Both writes take the `module.enable` permission, the same one that turns modules on and off. Both are reversible: switching testing off puts the project back on general versions from its next update, and a module's "general only" mark is kept if testing is switched off and on again. The choice is stored in the project's own settings (`modules.tester`, `modules.general_only`); anything that cannot be read counts as not testing. `testsVersionsOf` in `src/bongos/module-testing.js` is the one answer to "does this project get the tester version of this module", which installs and updates use (task 1003805). The hall's Modules tab has the switch in its "Testing new versions" group, and the general-only control in each module's details.

package/docs/page-readings.json

@@ -50,7 +50,7 @@
       "surface": "builders",
       "title": "Blockers",
       "states_read": ["default"],
-      "files_hash": "be570fb08f33cde4",
+      "files_hash": "f5bf3c2244e3fbdd",
       "reading_hash": "2b3fd98b7ba1bff6",
       "counts": {"lines":28,"placed":12,"shared":8,"unplaced":8},
       "omitted": {"data":9},
@@ -1141,7 +1141,7 @@
       "surface": "builders",
       "title": "The sky",
       "states_read": ["tour","sky","wells","ledger","full","full-back","out"],
-      "files_hash": "2dce7dcbe58dcf5d",
+      "files_hash": "5f2c2f22c2fc9dc6",
       "reading_hash": "b83eb8c92a23fd3c",
       "counts": {"lines":47,"placed":23,"shared":7,"unplaced":17},
       "omitted": {"data":8},
@@ -1201,7 +1201,7 @@
       "surface": "builders",
       "title": "Goals",
       "states_read": ["list","filters","cards","detail","dependencies","out","uat","uat-panel","author"],
-      "files_hash": "ca38b540196d951b",
+      "files_hash": "4f6f4eb5e9f00412",
       "reading_hash": "7aa1a6e1946dc077",
       "counts": {"lines":156,"placed":52,"shared":8,"unplaced":96},
       "omitted": {"data":268},
@@ -1326,7 +1326,7 @@
         {"key":"L0117","section":"Goals","text":"Version","placement":"unplaced"},
         {"key":"L0118","section":"Goals","text":"— any —","placement":"placed","file":"modules/hall-ui/public/hall-widgets.js","line":173,"string_id":"729d4ce22b99","occurrences":2},
         {"key":"L0119","section":"Goals","text":"Status","placement":"unplaced"},
-        {"key":"L0120","section":"Goals","text":"open","placement":"placed","file":"modules/hall-ui/public/approval-queue.js","line":164,"string_id":"9a803563c948","occurrences":16},
+        {"key":"L0120","section":"Goals","text":"open","placement":"placed","file":"modules/hall-ui/public/approval-queue.js","line":166,"string_id":"9a803563c948","occurrences":16},
         {"key":"L0121","section":"Goals","text":"Sort","placement":"unplaced"},
         {"key":"L0122","section":"Goals","text":"priority (version order)","placement":"unplaced"},
         {"key":"L0123","section":"Goals","text":"newest first","placement":"unplaced"},
@@ -1439,7 +1439,7 @@
       "surface": "builders",
       "title": "Builders' Hall",
       "states_read": ["home","standing","standing-karma","out","founding","genesis","decide","demo"],
-      "files_hash": "eb1e8c3a4621add8",
+      "files_hash": "e32eb071e46bec38",
       "reading_hash": "42dbab4f3948fe66",
       "counts": {"lines":117,"placed":63,"shared":10,"unplaced":44},
       "omitted": {"data":101},
@@ -1569,7 +1569,7 @@
       "surface": "builders",
       "title": "Idea",
       "states_read": ["record","missing"],
-      "files_hash": "5b7084faad7f4ded",
+      "files_hash": "7456ea53572161d4",
       "reading_hash": "c9f9a84ab844c704",
       "counts": {"lines":52,"placed":23,"shared":9,"unplaced":20},
       "omitted": {"data":18},
@@ -1609,7 +1609,7 @@
         {"key":"L0032","section":"Everything that fell out of it","text":"{…} and {…}","placement":"placed","file":"modules/hall-ui/public/profile-nudge.js","line":77,"string_id":"56b00d773232","state":"missing"},
         {"key":"L0033","section":"footer","text":"Sign out","placement":"shared","file":"modules/hall-ui/public/shell.js","line":520,"string_id":"1883ad7f07fc","state":"missing","occurrences":2},
         {"key":"L0034","section":"Everything that fell out of it","text":"Builders hall · this session","placement":"placed","file":"modules/hall-ui/public/craft-mode.js","line":674,"string_id":"fd1903586e8e","state":"missing"},
-        {"key":"L0035","section":"12A changelog every builder can read","text":"open","placement":"placed","file":"modules/hall-ui/public/approval-queue.js","line":164,"string_id":"9a803563c948"},
+        {"key":"L0035","section":"12A changelog every builder can read","text":"open","placement":"placed","file":"modules/hall-ui/public/approval-queue.js","line":166,"string_id":"9a803563c948"},
         {"key":"L0036","section":"12A changelog every builder can read","text":"had the thought · {…}","placement":"unplaced"},
         {"key":"L0037","section":"12A changelog every builder can read","text":"{…}/100 complete","placement":"placed","file":"modules/hall-ui/public/idea.js","line":87,"string_id":"c43050b62297"},
         {"key":"L0038","section":"12A changelog every builder can read","text":"{…} and {…}","placement":"placed","file":"modules/hall-ui/public/profile-nudge.js","line":77,"string_id":"56b00d773232","occurrences":4},
@@ -1719,7 +1719,7 @@
       "surface": "builders",
       "title": "Modules",
       "states_read": ["browse","search-empty","lens","details","out"],
-      "files_hash": "ce529d9f575606e5",
+      "files_hash": "6dc5460706966349",
       "reading_hash": "a17f501b72321b7b",
       "counts": {"lines":62,"placed":29,"shared":8,"unplaced":25},
       "omitted": {"data":106},
@@ -2325,7 +2325,7 @@
       "surface": "builders",
       "title": "Profile",
       "states_read": ["own","works","activity","activity-artist","activity-ideator","other","artist","ideator","out"],
-      "files_hash": "deca408267a32df9",
+      "files_hash": "82a35b34a5115325",
       "reading_hash": "8088d60a4e93bb91",
       "counts": {"lines":196,"placed":70,"shared":8,"unplaced":118},
       "omitted": {"data":197},
@@ -2721,7 +2721,7 @@
       "surface": "builders",
       "title": "Builders",
       "states_read": ["roster","artists","search","out"],
-      "files_hash": "38086c3c20a8cd6d",
+      "files_hash": "88e2f7aca1994346",
       "reading_hash": "5c7e507e9600e59c",
       "counts": {"lines":47,"placed":13,"shared":8,"unplaced":26},
       "omitted": {"data":60},
@@ -2781,7 +2781,7 @@
       "surface": "builders",
       "title": "Sessions",
       "states_read": ["ledger","out"],
-      "files_hash": "b5f59d12801a60d9",
+      "files_hash": "cd14ebd4d35e16bf",
       "reading_hash": "5c18aaed7f2e8e0f",
       "counts": {"lines":331,"placed":21,"shared":8,"unplaced":302},
       "omitted": {"data":303},
@@ -3978,7 +3978,7 @@
       "surface": "builders",
       "title": "Studio",
       "states_read": ["home","out","flow","flow-pick","status","words-offer","words-edit","reviews","full","full-back","approvals","approvals-before","approvals-phone","approvals-note","approved","sent-back"],
-      "files_hash": "bd33b2df51b981d4",
+      "files_hash": "a072c3fdf156d639",
       "reading_hash": "55e9090fdd96a979",
       "counts": {"lines":30,"placed":16,"shared":6,"unplaced":8},
       "omitted": {"data":4},
@@ -4021,7 +4021,7 @@
       "surface": "builders",
       "title": "Task",
       "states_read": ["record","from-goal","missing"],
-      "files_hash": "8200eeef7a793659",
+      "files_hash": "5535b9790b56985e",
       "reading_hash": "3e80d092af1d7e2c",
       "counts": {"lines":132,"placed":35,"shared":10,"unplaced":87},
       "omitted": {"data":75},
@@ -4166,7 +4166,7 @@
       "surface": "builders",
       "title": "Your thinking",
       "states_read": ["tour","sky","list","simple","full-idea","full","full-list","full-back","out"],
-      "files_hash": "f8061c0641b23730",
+      "files_hash": "a7b04635b758d676",
       "reading_hash": "532805c03c3a8a4c",
       "counts": {"lines":48,"placed":23,"shared":7,"unplaced":18},
       "omitted": {"data":8},
@@ -4263,7 +4263,7 @@
       "surface": "builders",
       "title": "Watch",
       "states_read": ["board","out","operations","security"],
-      "files_hash": "c3baec6343c9867a",
+      "files_hash": "3980ad82a03bc297",
       "reading_hash": "d698ac62c904aa65",
       "counts": {"lines":88,"placed":38,"shared":8,"unplaced":42},
       "omitted": {"data":166},
@@ -4364,7 +4364,7 @@
       "surface": "builders",
       "title": "Tasks",
       "states_read": ["ready","in-progress","backlog","completed","filters-open","out","founding"],
-      "files_hash": "2c4ec39162b32427",
+      "files_hash": "1dcb08711f496752",
       "reading_hash": "a425c7a090d8bf03",
       "counts": {"lines":125,"placed":54,"shared":8,"unplaced":63},
       "omitted": {"data":316},

package/modules/copy-desk/page-status.js

@@ -800,8 +800,40 @@ function approvalFor({ page, pageId, rounds, round, builderId, isArtist }) {
   };
 }
 
+// The submitted-but-unapplied rounds, oldest first, in the in-flight list's
+// own words (task 1004662's IN_FLIGHT.submitted), with how long each waited.
+function notAppliedFor({ byPage, invById, now }) {
+  const nowMs = now instanceof Date ? now.getTime() : Number(now);
+  const rows = [];
+  for (const [pageId, rounds] of byPage) {
+    const open = rounds.filter((r) => !TERMINAL.has(r.status));
+    const round = open.length ? open[open.length - 1] : null;
+    if (!round || roundState(round) !== 'submitted') continue;
+    const page = invById.get(pageId) || null;
+    const b = pages.parseBatchBlock(round.description);
+    const since = iso((b.ok && b.batch.submitted_at) || round.updated_at || round.created_at);
+    const ms = since ? nowMs - Date.parse(since) : null;
+    const surface = page ? page.surface : String(pageId).split(':')[0];
+    rows.push({
+      page_id: pageId,
+      surface,
+      surface_word: surfaceWord(surface),
+      title: page ? (page.title || page.id) : pageId,
+      path: page ? page.path || null : null,
+      task_id: String(round.id),
+      state: 'submitted',
+      state_word: IN_FLIGHT.submitted.word,
+      since,
+      waited_days: ms == null ? null : Math.max(0, Math.floor(ms / DAY_MS)),
+      line_count: b.ok ? b.batch.lines.length : 0,
+    });
+  }
+  rows.sort((a, b) => String(a.since || '').localeCompare(String(b.since || '')) || Number(a.task_id) - Number(b.task_id));
+  return { not_applied: rows, not_applied_count: rows.length };
+}
+
 // GET /copy-desk/approvals — the queue, "1 of 2".
-function composeApprovals({ inventory, tasks, builderId, crafts }) {
+function composeApprovals({ inventory, tasks, builderId, crafts, now = Date.now() }) {
   const byPage = indexRounds(tasks);
   const invById = new Map(((inventory && inventory.pages) || []).filter((p) => p && p.id).map((p) => [p.id, p]));
   const isArtist = (crafts || []).includes('artist');
@@ -825,6 +857,11 @@ function composeApprovals({ inventory, tasks, builderId, crafts }) {
     is_artist: isArtist,
     count: queue.length,
     queue,
+    // task 1004765: rounds SUBMITTED but not yet applied. They sat unseen for
+    // weeks because this queue listed only applied rounds; they are listed
+    // apart, never in `queue`, so nothing can approve or send back a page that
+    // has not been applied, and `count` (the studio door) stays reviewable only.
+    ...notAppliedFor({ byPage, invById, now }),
   };
 }
 

package/modules/hall-ui/public/approval-queue-lib.js

@@ -130,6 +130,20 @@
     };
   }
 
+  // task 1004765: a submitted round nobody has applied yet. Listed under the
+  // reviewable pages, marked so, with how long it has waited; it carries no
+  // verb, since there is no applied page to judge.
+  function notAppliedRow(r) {
+    const d = r && r.waited_days;
+    const waited = d == null ? '' : (d === 0 ? 'today' : `${d} day${d === 1 ? '' : 's'}`);
+    return {
+      task_id: String(r.task_id),
+      title: displayTitle(r),
+      mark: 'not applied yet',
+      line: `${r.state_word}${waited ? ` · waiting ${waited}` : ''}. The words are waiting to be applied, so there is nothing to approve yet.`,
+    };
+  }
+
   function noteProblem(note) {
     const n = String(note == null ? '' : note).replace(/\s+/g, ' ').trim();
     if (n.length < MIN_NOTE) return `Say what needs changing in at least ${MIN_NOTE} characters.`;
@@ -157,7 +171,7 @@
   const api = {
     PHASES, DEVICES, MODES, MIN_NOTE, MAX_NOTE,
     slotName, renderIndex, pickRender, hasAnyRender, displayTitle,
-    eyebrow, subLine, doorDetail, refusedWords, lineRow, noteProblem, approvedLine, sentBackLine,
+    eyebrow, subLine, doorDetail, refusedWords, lineRow, notAppliedRow, noteProblem, approvedLine, sentBackLine,
   };
   if (typeof window !== 'undefined') window.OTBApprovalQueueLib = api;
   if (typeof module !== 'undefined' && module.exports) module.exports = api;

package/modules/hall-ui/public/approval-queue.css

@@ -230,6 +230,15 @@ html.aq-open { overflow: hidden; }
 .aq-line__pinned { font-size: 13px; line-height: 1.45; color: var(--aq-link); overflow-wrap: anywhere; }
 .aq-lines__none { margin: 0; font-size: 14px; color: var(--aq-ink-faint); }
 
+/* task 1004765: submitted rounds not applied yet, below the reviewable page. */
+.aq-pending { padding: 14px 20px 18px; border-top: 1px solid var(--aq-rule); }
+.aq-pending__h { margin: 0 0 8px; font-size: 13px; font-weight: 600; color: var(--aq-ink-faint); }
+.aq-pending__list { list-style: none; margin: 0; padding: 0; display: flex; flex-direction: column; gap: 8px; }
+.aq-pending__row { display: flex; flex-wrap: wrap; gap: 4px 8px; font-size: 14px; }
+.aq-pending__title { font-weight: 600; }
+.aq-pending__mark { font-family: var(--aq-mono); font-size: 11px; letter-spacing: .08em; text-transform: uppercase; color: var(--aq-link); }
+.aq-pending__line { flex-basis: 100%; font-size: 13px; color: var(--aq-ink-faint); overflow-wrap: anywhere; }
+
 /* The decision row. */
 .aq-foot {
   display: flex;

package/modules/hall-ui/public/approval-queue.js

@@ -105,6 +105,7 @@
       state.loadError = null;
       state.loaded = true;
       state.queue = (res.data && res.data.queue) || [];
+      state.notApplied = (res.data && res.data.not_applied) || [];
       if (state.idx >= state.queue.length) state.idx = 0;
       tellCount();
     }
@@ -183,6 +184,19 @@
       return `<div class="aq-foot">${approve}${note}<span class="aq-hint">highlighted: the words you changed</span></div>${err}`;
     }
 
+    // task 1004765: submitted rounds nobody has applied, below the reviewable
+    // pages. The server lists them apart; the panel only words them, and gives
+    // them no approve or send-back verb.
+    function notAppliedHtml() {
+      const rows = (state.notApplied || []).map((r) => L.notAppliedRow(r));
+      if (!rows.length) return '';
+      return `<section class="aq-pending" aria-label="Submitted, not applied yet">
+          <h3 class="aq-pending__h">Submitted, not applied yet (${rows.length})</h3>
+          <ul class="aq-pending__list">${rows.map((r) => `
+            <li class="aq-pending__row" data-task="${esc(r.task_id)}"><span class="aq-pending__title">${esc(r.title)}</span> <span class="aq-pending__mark">${esc(r.mark)}</span><span class="aq-pending__line">${esc(r.line)}</span></li>`).join('')}</ul>
+        </section>`;
+    }
+
     function paint() {
       if (!state.open) return;
       const keepNote = document.getElementById('aq-note');
@@ -193,7 +207,8 @@
         const words = state.loadError || (state.loaded ? 'Nothing is waiting for you. A page appears here once /tweak has applied it and its grade has passed.' : 'Opening the Approval queue…');
         dialog.innerHTML = `
           <div class="aq-head"><span class="aq-eyebrow">approval queue</span><h2 class="aq-title" id="aq-title">${state.loaded ? 'Nothing waiting' : 'Approval queue'}</h2><span class="aq-head__gap"></span>${close}</div>
-          <div class="aq-body aq-body--empty"><div class="aq-frame"><p class="aq-frame__note">${esc(words)}</p></div></div>`;
+          <div class="aq-body aq-body--empty"><div class="aq-frame"><p class="aq-frame__note">${esc(words)}</p></div></div>
+          ${notAppliedHtml()}`;
         return;
       }
       loadRenders(item);
@@ -212,7 +227,8 @@
             ${linesHtml(item)}
           </aside>
         </div>
-        ${footHtml(item)}`;
+        ${footHtml(item)}
+        ${notAppliedHtml()}`;
       const note = document.getElementById('aq-note');
       if (note) { note.value = noteValue; }
     }

package/modules/hall-ui/public/brand-apply.js

@@ -91,7 +91,7 @@
   // estimate with a "<Currency> Est." label; settled → the actual ledger payout
   // with a plain "<Currency>" label.
   //
-  // task 1002469 — on a cost-plus-only instance that second case is a LIE. The
+  // task 1002469 (superseded by task 1004781) — on a cost-plus-only instance that second case was a LIE. The
   // per-task estimate is never paid, so credits_awarded is 0 on every shipped
   // task, taskCreditValue's zero-fallback hands back the estimate, and the plain
   // label then presents that estimate as the actual payout. So the "Est."
@@ -105,18 +105,25 @@
   function taskCreditValue(task) {
     if (!task) return null;
     if (task.status === 'abandoned' || task.merged_into_task_id != null) return 0;
+    // task 1004781 — a settled task is worth what the ledger PAID for it and
+    // nothing else. It used to fall back to the estimate when the ledger held
+    // 0, and on a cost-plus-only instance that is every shipped task, so goal
+    // totals summed guesses as money (a goal read 3,373 for a builder paid
+    // 776). A settled 0 is the truth: the task itself paid nothing. Only
+    // unsettled work carries the estimate, and taskCreditLabel marks it "Est.".
+    // The decision the task asked for: estimates are KEPT, for live work only —
+    // they still rank claimable work (task-lib's best-value sort) and tell a
+    // builder roughly how big a task is — but they never reach a total.
     if (taskSettled(task)) {
       var awarded = Number(task.credits_awarded);
-      // Settled with a real ledger payout → show it. If a settled task has no
-      // ledger record (0 — legacy pre-ledger tasks, or reward-exempt), fall back
-      // to the estimate rather than showing a misleading "0".
-      if (isFinite(awarded) && awarded > 0) return awarded;
+      return isFinite(awarded) && awarded > 0 ? awarded : 0;
     }
     return task.credits_reward;
   }
   function taskCreditLabel(task, opts) {
+    // A settled figure is now always the actual payout (see taskCreditValue),
+    // so it takes the plain label on every reward mode; only an estimate is "Est.".
     var base = currencyLabel(opts);
-    if (!perTaskCredit()) return base + ' Est.';
     return taskSettled(task) ? base : base + ' Est.';
   }
 

package/modules/hall-ui/public/sky-data.js

@@ -142,15 +142,16 @@
   // The prototype read a snapshot (docs/design/mocks/sky/snapshot.js). Each
   // adaptor maps one real read onto the same field names, so the world below
   // is the prototype's world line for line.
+  const SETTLED = new Set(['confirmed', 'shipped']);
   function goalFromSky(g) {
     return {
       id: String(g.goal_id), title: g.title || 'A private goal', subtitle: g.subtitle || null, version: g.version_id, status: g.status,
       taskTotal: Number(g.task_total) || 0, taskDone: Number(g.task_done) || 0, createdAt: g.created_at || null,
       crit: (g.criteria || []).map((c) => ({ id: String(c.id), text: c.criterion_md || '', satisfied: c.satisfied === true, tasks: (c.task_ids || []).map(String), pendingReview: c.pending_review === true, uat: c.uat_state || null, uatLabel: uatLabel(c.uat_state), backendOnly: c.backend_only === true })),
-      // task 1004782 — abandoned (or merged-away) work carries no credits: it is listed, at zero, and never adds to a
-      // well's mass or a ring. A merged-away task is always abandoned (tasks.merged_into_task_id is CHECK-constrained to
-      // abandoned rows), so the status test covers both; /sky does not send the pointer.
-      tasks: (g.tasks || []).map((t) => ({ id: String(t.id), title: t.title || 'a task', status: t.status, credits: t.status === 'abandoned' || t.credits_reward == null ? 0 : Number(t.credits_reward), createdAt: t.created_on || null, sourceRef: t.idea_id != null ? String(t.idea_id) : null })),
+      // `credits` is what the ledger PAID — settled work only (task 1004781) — so a well's mass and a ring's weight are
+      // money paid. Live work carries its guess as `estimate`, shown "Est." and never added to a total. Abandoned (or
+      // merged-away, which is always abandoned — CHECK-constrained) work has neither (task 1004782).
+      tasks: (g.tasks || []).map((t) => ({ id: String(t.id), title: t.title || 'a task', status: t.status, credits: SETTLED.has(t.status) ? Number(t.credits_awarded) || 0 : 0, estimate: SETTLED.has(t.status) || t.status === 'abandoned' || t.credits_reward == null ? null : Number(t.credits_reward), createdAt: t.created_on || null, sourceRef: t.idea_id != null ? String(t.idea_id) : null })),
     };
   }
   const taskRow = (t) => ({ id: String(t.task_id ?? t.id), title: t.title || 'a task', status: t.status, goalId: t.goal_id != null ? String(t.goal_id) : null, origin: t.origin || null });

package/modules/hall-ui/public/sky-panel.js

@@ -98,7 +98,7 @@
       <div class="facts">
         ${g ? fact('well', esc(g.title)) : ''}
         ${crit ? fact('gates the ring', esc(plain(crit.text).slice(0, 90))) : g ? fact('gates the ring', 'none — outer orbit') : ''}
-        ${t.credits != null ? fact('credits', String(t.credits), true) : ''}
+        ${t.estimate != null ? fact('credits', `${t.estimate} <span class="lbl">Est.</span>`, true) : t.credits != null ? fact('credits paid', String(t.credits), true) : ''}
         ${t.createdAt ? fact('born', esc(t.createdAt)) : ''}
         ${t.sourceRef ? fact('born of a thought', `<a href="${HALL}/idea/${esc(t.sourceRef)}" target="_blank" rel="noopener">${esc(ideaTitle(t.sourceRef))}</a>`) : ''}
       </div>

package/modules/lifecycle/sky.js

@@ -44,6 +44,7 @@ const goalGraph = require('./goal-graph');
 const goalCategory = require('./goal-category');
 const { isPendingReview, uatState, uatReadColumnsSql } = require('./done-when');
 const { TERMINAL_STATUSES } = require('./task-status');
+const { CREDITS_AWARDED_SUBQUERY } = require('./db-tasks');
 
 const SKY_CACHE_TTL_MS = 60 * 1000;
 
@@ -92,9 +93,12 @@ async function skyRows({ versionId = null, viewerId = null, viewerIsArchon = fal
       // Compact by construction: no description, no value summary, no notes.
       // created_at is truncated to the DAY in SQL so the payload carries ten
       // characters per task, not twenty-seven, and the clock draws by day anyway.
+      // credits_awarded (task 1004781) is the shared settled-only payout subquery; its
+      // per-row lookup rides idx_credit_log_task (core_189), never a scan of the log.
       `SELECT t.id::text AS id, t.goal_id::text AS goal_id, t.title, t.status,
               t.credits_reward, t.created_at::date::text AS created_on,
-              CASE WHEN t.source_ref ~ '^idea:[0-9]+$' THEN substring(t.source_ref from 6) END AS idea_id
+              CASE WHEN t.source_ref ~ '^idea:[0-9]+$' THEN substring(t.source_ref from 6) END AS idea_id,
+              ${CREDITS_AWARDED_SUBQUERY} AS credits_awarded
          FROM tasks t
          JOIN goals g ON g.id = t.goal_id
         WHERE ($1::text IS NULL OR g.version_id = $1)
@@ -189,6 +193,9 @@ function shapeSky({ goals = [], edges = [], extras = [], criteria = [], tasks =
       title: visible ? t.title : null,
       status: t.status,
       credits_reward: t.credits_reward == null ? null : Number(t.credits_reward),
+      // task 1004781 — what the ledger actually paid (0 until settled), so the
+      // sky sizes a well by money paid, never by the estimate.
+      credits_awarded: t.credits_awarded == null ? 0 : Number(t.credits_awarded),
       created_on: t.created_on || null,
       idea_id: t.idea_id != null ? String(t.idea_id) : null,
     });

package/package-lock.json

@@ -1,12 +1,12 @@
 {
   "name": "@bongos/core",
-  "version": "1.21.169",
+  "version": "1.21.171",
   "lockfileVersion": 3,
   "requires": true,
   "packages": {
     "": {
       "name": "@bongos/core",
-      "version": "1.21.169",
+      "version": "1.21.171",
       "license": "AGPL-3.0-or-later",
       "dependencies": {
         "express": "^4.21.2",

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@bongos/core",
-  "version": "1.21.169",
+  "version": "1.21.171",
   "description": "Cloud Bongos — the AI-first build platform core (GDS + platform surfaces + module system), installed as a versioned dependency (ADR 0108).",
   "license": "AGPL-3.0-or-later",
   "main": "src/platform-server.js",

package/release-notes.json

@@ -10084,5 +10084,21 @@
       "id": "1003805",
       "text": "Projects that test new module versions now actually get them first, and everyone else only gets versions that have passed testing. Turning testing off keeps what you have and moves you to the tested versions from your next upd"
     }
+  ],
+  "1.21.170": [
+    {
+      "id": "1004765",
+      "text": "The Approval queue now also shows pages an artist submitted that nobody has applied yet, under the pages ready to review. Each is marked not applied yet and says how many days it has waited. They have no approve or send back b"
+    },
+    {
+      "id": "1003806",
+      "text": "When a new module version is being tested, how it went for the testing projects now counts toward its quality score — and a handful of friendly testers can no longer outscore hundreds of real ones."
+    }
+  ],
+  "1.21.171": [
+    {
+      "id": "1004781",
+      "text": "Credit figures in the hall now show money that was actually paid. A finished task shows its real payout instead of the old guess, so goal totals, member stats and the goal map stop adding up estimates as if they were earnings."
+    }
   ]
 }

package/scripts/gds/module-assess-installs.js

@@ -1,6 +1,6 @@
 #!/usr/bin/env node
-// scripts/gds/module-assess-installs.js — the Install and Reliability parts of a
-// module's assessment, worked out from what real installs report (task 1003793,
+// scripts/gds/module-assess-installs.js — the Install, Reliability and Tester
+// feedback parts of a module's assessment, worked out from what real installs report (task 1003793,
 // ADR 0343 D1; reports from core_273, signal rows into core_265).
 //
 // WHY these two parts. Tests and Docs are things an author writes; Install and
@@ -17,6 +17,21 @@
 //                  reported), over the last ERROR_DAY_WINDOW days.
 // No reports → no_data (ADR 0343 D3: left out of the average, never a zero).
 //
+// Tester feedback (task 1003806, ADR 0344 D5) is the same reports, narrowed to
+// the version's TESTER WINDOW: from the first time it went on the tester channel
+// to the first time it left it (promoted or withdrawn), or now while it is still
+// there. Only tester projects are served a tester version (ADR 0344 D1), so a
+// report inside that window is a tester's. A tester it went well for is one whose
+// install did not fail and who reported no runtime errors in the window.
+// Sample size has to weigh — two friendly testers must not outrank two hundred
+// real ones when a buyer sorts by this part — so the count is blended with
+// PRIOR_TESTERS imaginary testers it went half well for:
+//   Tester feedback = round(100 × (went well + PRIOR_TESTERS/2) / (testers + PRIOR_TESTERS)).
+// 3 of 3 → 62, 10 of 10 → 75, 200 of 200 → 98. MIN_HOLDERS applies here too.
+// An install report is one row per holder, re-stamped on each attempt, so a
+// tester who reinstalls after promotion leaves the window; error counts are per
+// day, so they are placed by their day, not by when they were sent.
+//
 // What the reports may contain is fixed by the owner's privacy posture and
 // enforced at intake (src/bongos/module-usage-reports.js): counts and short codes
 // only, and nothing from a holder who switched reporting off.
@@ -29,6 +44,7 @@
 const { ERROR_DAY_WINDOW } = require('../../src/bongos/module-usage-reports');
 
 const MIN_HOLDERS = 3;
+const PRIOR_TESTERS = 10;
 
 // The not-yet-scored outcomes, shared by both parts. Pure.
 function notYet(holders, what) {
@@ -67,8 +83,41 @@ function assessReliability(days = []) {
   };
 }
 
+// Tester feedback from what testers reported in the tester window: `installs` is
+// { holder_kind, holder_ref, ok, error_code } per holder, `days` the holder-days
+// as Reliability takes them. One verdict per tester, blended with the prior. Pure.
+function assessTesterFeedback({ installs = [], days = [] } = {}) {
+  const testers = new Map();
+  const of = (r) => {
+    const k = `${r.holder_kind}:${r.holder_ref}`;
+    if (!testers.has(k)) testers.set(k, { failed: false, errors: 0 });
+    return testers.get(k);
+  };
+  for (const r of installs) of(r).failed = !r.ok;
+  for (const d of days) of(d).errors += Number(d.error_count);
+  const all = [...testers.values()];
+  const n = all.length;
+  const wentWell = all.filter((t) => !t.failed && t.errors === 0).length;
+  const detail = {
+    testers: n, went_well: wentWell,
+    install_failed: all.filter((t) => t.failed).length,
+    with_errors: all.filter((t) => t.errors > 0).length,
+    errors: all.reduce((sum, t) => sum + t.errors, 0),
+    prior_testers: PRIOR_TESTERS,
+  };
+  if (n < MIN_HOLDERS) return { ...notYet(n, 'from the tester channel'), detail };
+  return {
+    outcome: 'scored',
+    score: Math.round((100 * (wentWell + PRIOR_TESTERS / 2)) / (n + PRIOR_TESTERS)),
+    sample_size: n,
+    reason: `it went well for ${wentWell} of ${n} testers (weighed against ${PRIOR_TESTERS} neutral ones, so a small sample cannot score high)`,
+    detail,
+  };
+}
+
 // The author's own reports never count. Shared by both reads.
-const NOT_AUTHOR = `NOT (r.holder_kind = 'builder' AND r.holder_ref = m.author_id::text)`;
+const notAuthor = (a) => `NOT (${a}.holder_kind = 'builder' AND ${a}.holder_ref = m.author_id::text)`;
+const NOT_AUTHOR = notAuthor('r');
 
 async function readReports(versionId, pool) {
   const { rows: latest } = await pool.query(
@@ -87,26 +136,62 @@ async function readReports(versionId, pool) {
   return { latest, days };
 }
 
+// The version's tester window, as SQL over its channel history (core_276): `t.start`
+// is when it first went on the tester channel, `t.stop` when it next left it (null
+// while it is still there). A version that was never a tester version has no row.
+const TESTER_WINDOW = `(SELECT s.start,
+          (SELECT min(c2.set_at) FROM store_module_version_channels c2
+            WHERE c2.version_id = $1 AND c2.channel <> 'tester' AND c2.set_at > s.start) AS stop
+     FROM (SELECT min(c1.set_at) AS start FROM store_module_version_channels c1
+            WHERE c1.version_id = $1 AND c1.channel = 'tester') s
+    WHERE s.start IS NOT NULL) t`;
+
+async function readTesterReports(versionId, pool) {
+  const { rows: installs } = await pool.query(
+    `SELECT ir.holder_kind, ir.holder_ref, ir.ok, ir.error_code
+       FROM module_install_reports ir
+       JOIN store_module_versions v ON v.id = ir.version_id
+       JOIN store_modules m ON m.module_key = v.module_key
+       CROSS JOIN ${TESTER_WINDOW}
+      WHERE ir.version_id = $1 AND ir.reported_at >= t.start AND (t.stop IS NULL OR ir.reported_at < t.stop)
+        AND ${notAuthor('ir')}`, [versionId]);
+  const { rows: days } = await pool.query(
+    `SELECT ec.holder_kind, ec.holder_ref, ec.error_count
+       FROM module_runtime_error_counts ec
+       JOIN store_module_versions v ON v.id = ec.version_id
+       JOIN store_modules m ON m.module_key = v.module_key
+       CROSS JOIN ${TESTER_WINDOW}
+      WHERE ec.version_id = $1 AND ec.day >= t.start::date AND (t.stop IS NULL OR ec.day <= t.stop::date)
+        AND ${notAuthor('ec')}`, [versionId]);
+  return { installs, days };
+}
+
 const sameResult = (cur, s) => !!cur && cur.outcome === s.outcome
   && (cur.score == null ? null : Number(cur.score)) === s.score
   && Number(cur.sample_size) === s.sample_size;
 
-// Re-score Install and Reliability for one version from its reports, append a
+const PARTS = ['install', 'reliability', 'tester_feedback'];
+
+// Re-score Install, Reliability and Tester feedback for one version from its reports, append a
 // signal row for each part whose result moved, then re-compose the overall score
 // (module-assess-score.js). Resolves { ok, written: [part…], score }.
 async function recordInstallSignals(versionId, { db } = {}) {
   const pool = db || require('../../src/bongos/pool').pool;
   const { latest, days } = await readReports(versionId, pool);
-  const next = { install: assessInstall(latest), reliability: assessReliability(days) };
+  const next = {
+    install: assessInstall(latest),
+    reliability: assessReliability(days),
+    tester_feedback: assessTesterFeedback(await readTesterReports(versionId, pool)),
+  };
 
   const { rows: current } = await pool.query(
     `SELECT DISTINCT ON (part) part, outcome, score, sample_size
-       FROM module_assessment_signals WHERE version_id = $1 AND part IN ('install', 'reliability')
+       FROM module_assessment_signals WHERE version_id = $1 AND part IN ('install', 'reliability', 'tester_feedback')
       ORDER BY part, measured_at DESC, id DESC`, [versionId]);
   const byPart = Object.fromEntries(current.map((r) => [r.part, r]));
 
   const written = [];
-  for (const part of ['install', 'reliability']) {
+  for (const part of PARTS) {
     const s = next[part];
     if (sameResult(byPart[part], s)) continue;
     await pool.query(
@@ -128,7 +213,7 @@ async function main(argv) {
     return 2;
   }
   const res = await recordInstallSignals(id);
-  for (const part of ['install', 'reliability']) {
+  for (const part of PARTS) {
     const s = res.signals[part];
     console.log(`${part}: ${s.outcome}${s.score != null ? ` ${s.score}/100` : ''} — ${s.reason}${res.written.includes(part) ? '' : ' (unchanged)'}`);
   }
@@ -140,4 +225,4 @@ if (require.main === module) {
   main(process.argv.slice(2)).then((code) => { process.exitCode = code; }, (e) => { console.error(e.stack || e.message); process.exitCode = 1; });
 }
 
-module.exports = { MIN_HOLDERS, assessInstall, assessReliability, recordInstallSignals };
+module.exports = { MIN_HOLDERS, PRIOR_TESTERS, assessInstall, assessReliability, assessTesterFeedback, recordInstallSignals };