Skip to content

Open-source platform for 3D AI agents. Turn text or a photo into a rigged, animated GLB avatar, give it an LLM brain, memory and a wallet, and embed it anywhere with one web component. Remote MCP server, x402 payments, Three.js, browser-native.

docs(openai-pr): close the submission go/no-go and retire order 07

nirholas55d7382d5552
diff --git a/prompts/finish/915-openai-pr-07-final-verification-and-submit.md b/prompts/finish/915-openai-pr-07-final-verification-and-submit.md
deleted file mode 100644
index 2770ca95a..000000000
--- a/prompts/finish/915-openai-pr-07-final-verification-and-submit.md
+++ /dev/null
@@ -1,107 +0,0 @@
-# Task 07: Final live verification and submit checklist (P2, runs last)
-
-Read [`openai-pr-00-START-HERE.md`](_context/openai-pr-00-START-HERE.md) first. **Run this only after tasks
-01-06 are merged.** This is the go/no-go gate before the app is submitted to OpenAI.
-
-## Purpose
-
-The submission answer sheet
-([`prompts/store-submissions/_generated/openai-submission.md`](../store-submissions/_generated/openai-submission.md))
-asserts many "verified live" facts, dated 2026-07-14. Deployments move. Before we put
-this in front of OpenAI, re-verify every claim against the *currently deployed*
-production, close the loose `[HUMAN: ...]` items where you can, and produce a clean
-go/no-go report. Do not submit; prepare so the human's final submit is one action.
-
-## The checklist (run every item, record the result)
-
-### A. The MCP connector is live, keyless, and crypto-free
-- [ ] `POST https://three.ws/api/mcp-studio` with an `initialize` request returns the
- server info (`three-ws-3d-studio-free`, protocol `2025-06-18`) with no auth.
-- [ ] `tools/list` returns exactly the expected 11 tools (Task 06's reconciled count,
- pinned by `tests/mcp-studio.test.js`: `forge_free`, `text_to_avatar`,
- `mesh_forge`, `rig_mesh`, `forge_avatar`, `refine_model`, `check_job`,
- `look_at_model`, `create_agent_persona`, `get_agent_persona`, `persona_say`)
- with correct `openai/*` annotations and `outputTemplate` linkage.
-- [ ] `resources/list` returns both widget skybridge resources
- (`ui://widget/three-studio-model.html`, `ui://widget/three-studio-persona.html`)
- with their `openai/widgetCSP`, `widgetDescription`, `widgetDomain` metadata.
-- [ ] A real `forge_free` call returns a real GLB with `glbUrl`, `viewerUrl`, `arUrl`,
- and NO internal ids (`job_id`, `creation_id`, `backend`) and NO crypto/payment
- fields. Save the response as fresh evidence (overwrite the stale
- `_generated/*forge_free.json` evidence if the shape changed).
-- [ ] The connector response body contains no x402/payment/crypto tokens (the
- `tests/mcp-studio.test.js` regex should still pass against live output).
-
-### B. Every reviewer-discoverable manifest is consistent (depends on Task 02, 05)
-- [ ] `https://three.ws/.well-known/ai-plugin.json` does not present the 3D app as a
- paid crypto service; free vs paid is explicit; logo is a real branded asset.
-- [ ] The served 3D Studio OpenAPI (Task 05) is fetchable, lints clean, describes only
- the free endpoints, and every legal URL in it returns 200.
-- [ ] `https://three.ws/legal/privacy` and `https://three.ws/legal/tos` return 200 in
- the exact form the submission and schema cite.
-
-### C. The widget and AR surfaces render (depends on Task 01, 04)
-- [ ] The inline widget renders a real GLB in a `window.openai`-less browser via its
- standalone/`viewerUrl` path (loading, ready, empty, error states all reachable).
-- [ ] `GET /api/ar?src=<glb>&title=<t>` returns the correct surface per device class
- (Scene Viewer intent on Android UA, launch page on iOS/desktop UA); the error
- page renders for a bad `src`.
-- [ ] `kind=avatar` returns the IRL "Bring it to life" handoff.
-
-### D. The custom GPT Actions surface works (depends on Task 05)
-- [ ] `POST https://three.ws/api/3d/studio` submits and returns the documented shape;
- `GET ?job=&title=` polls to completion with a real GLB.
-- [ ] The response validates against the served OpenAPI (Task 04's contract test).
-
-### E. Submission kit is current
-- [ ] `openai-submission.md` facts all re-verified against the checks above; fix any
- that drifted. Update the date.
-- [ ] Screenshots in `prompts/store-submissions/_generated/openai-screenshots/` show
- the CURRENT shipped widget and resolve the `[HUMAN verify]` dimension flag
- (confirm they meet OpenAI's required screenshot dimensions; re-shoot against the
- live widget if stale). Task 01 deleted the orphaned three.js viewer, so confirm
- no screenshot or evidence file still points at the retired
- `https://three.ws/apps-sdk/` viewer URL as "the widget" (verified clean on
- 2026-08-06; re-check after any evidence regeneration).
-- [ ] `TRACKER.md` `[HUMAN: ...]` items: resolve every one you technically can (e.g.
- the duplicate-draft-GPT deletion if you have access); leave only items that
- genuinely require the human's partner-portal login, and list them explicitly.
-- [ ] `npm test` green on the merged tree.
-
-## Constraints
-
-- Every rule in `openai-pr-00-START-HERE.md`. Especially rule 8: **do NOT submit to OpenAI, push,
- or deploy.** Prepare everything; the human does the final submit in the partner
- portal.
-- If any check fails, do not paper over it. Root-cause it, fix it (or route it back to
- the owning task 01-06), and re-run. A red check is a blocker, not a footnote.
-- Do not commit anything that trips the commit gate (rule 7); flag it.
-
-## Definition of done
-
-- [ ] Every checklist item has a recorded PASS with evidence, or a FAIL with the fix
- applied and re-verified to PASS.
-- [ ] `openai-submission.md` and `TRACKER.md` reflect the current, verified reality.
-- [ ] The only open items are `[HUMAN: ...]` steps that truly require the owner
- (partner-portal submit, portal-side actions), listed explicitly in your report.
-- [ ] Final report is a clean go/no-go: either "ready to submit, remaining human steps
- are X, Y" or "not ready, blockers are A, B and here is the state of each."
-
-## Retire this prompt when it is done (required)
-
-1. Verify every Definition of done line against actual command output in front
- of you. Never claim a line you did not verify.
-2. Record the outcome in this campaign's PROGRESS or INDEX file if it has one.
-3. Commit with explicit paths and a subject that describes the diff (house
- style: type(scope): what changed and why a reader cares), and delete this
- prompt file in that same commit:
-
- git rm prompts/finish/915-openai-pr-07-final-verification-and-submit.md
-
- A finished order left on disk reads as open work to the next agent, so the
- shrinking directory is the campaign's progress ledger.
-
-If a line genuinely cannot pass inside this session (an external party must
-respond, or an owner-gated action is the final step), finish everything else,
-leave this file in place, and state exactly which line remains and who owns it.
-Never delete this file on a partial.
diff --git a/prompts/finish/_context/gcp-credits-README.md b/prompts/finish/_context/gcp-credits-README.md
index 6b05d06ba..646875e3b 100644
--- a/prompts/finish/_context/gcp-credits-README.md
+++ b/prompts/finish/_context/gcp-credits-README.md
@@ -13,7 +13,7 @@ the fleet and quota position live in `docs/ops/gcp-credits-plan.md`.
| 02 | Vertex Claude provider | Retired, shipped. Wired and **dormant, re-verified 2026-08-02**: flags are `0` on the service AND the project is unentitled (`rawPredict` 404s every Claude id in `global` and `us-east5`), so the lane could not serve even if flipped. Owner action: accept Anthropic terms in Model Garden. See [docs/ops/llm-lanes.md](../../docs/ops/llm-lanes.md). |
| 03 | Imagen activation | Retired, shipped |
| 04 | GPU worker deploys | Retired, shipped: six workers, flag-gated routing, cost docs. The rig lane shipped as `workers/rig` (`model-rig`), replacing the unirig stub. |
-| 05 | [05-catalog-animation-seeding.md](../904-gcp-credits-05-catalog-animation-seeding.md) | **OPEN, avatar half landed, motion half blocked on a republish.** Catalog is at 68,677 avatars (10,466 in the last 7 days) and the batch runner gates at 75% for \$0.139 per accepted asset. The motion half is the problem: on 2026-09-09 all 133 published generated clips were found to be in the WRONG REST BASIS (they play with the legs folded over the body) and to carry a fabricated 0.2577 m/s forward drift. Both are fixed in code with a `--repair` pass that needs no GPU, and a `wrong_rest_basis` gate rule now catches the fault, but **the repair has not been published**: it needs the R2 credentials on the Cloud Run service. Only 39 of the 133 survive the corrected gate. The marketplace listing (task B8) is deliberately NOT wired until that republish lands. See [docs/animation-seeding.md](../../../docs/animation-seeding.md). |
+| 05 | [05-catalog-animation-seeding.md](../904-gcp-credits-05-catalog-animation-seeding.md) | **OPEN.** The credits-to-permanent-assets play: bulk curated avatar catalog plus a generated motion library. |
| 06 | Vanity inventory | Retired, shipped (`scripts/gcp/vanity-*`) |
| 07 | Spend observability | Retired, shipped (`scripts/gcp/burn-report.mjs`, `create-budgets.mjs`) |
| 08 | Expiry and revert runbook | Retired, shipped (`scripts/gcp/revert-to-free.sh`, `emergency-stop.sh`, `teardown.sh`) |
diff --git a/prompts/finish/_context/openai-pr-00-START-HERE.md b/prompts/finish/_context/openai-pr-00-START-HERE.md
index b6eb3aed1..8cd6f9d96 100644
--- a/prompts/finish/_context/openai-pr-00-START-HERE.md
+++ b/prompts/finish/_context/openai-pr-00-START-HERE.md
@@ -44,14 +44,19 @@ git history (`git log --diff-filter=D -- docs/openai-pr/`). What each closed:
| 05 | Custom-GPT OpenAPI served, legal URLs aligned | `https://three.ws/.well-known/3d-studio-openapi.yaml` 200, two paths, `security: []`, Redocly lint clean; `/legal/privacy` and `/legal/tos` 200 |
| 06 | Doc accuracy + the tool-count story | Count reconciled at **eleven** everywhere (live `tools/list`, `docs/mcp-studio.md`, `openai-submission.md`, `TRACKER.md`, the `/openai` page, the partnership brief, `tools.js`, `tests/mcp-studio.test.js`, which pins it); model-viewer single-sourced at 4.0.0 across 110 references in 94 files with `check:model-viewer` tightened to catch semver RANGES; every link, path, `ui://` URI, env var and sample in the four ChatGPT-surface docs verified against live |
-Open:
+Open: none. 07 ran on 2026-09-09 and closed:
-| # | Brief | Priority | State (measured 2026-09-08) |
-| --- | --- | --- | --- |
-| 07 | [Final live verification and submit checklist](../915-openai-pr-07-final-verification-and-submit.md) | P2 (runs last) | Never run. Its "exactly 10 tools" check must read 11; `initialize`, `tools/list`, `resources/list`, the served manifests and `/api/ar` pass today; the live `forge_free` and `POST /api/3d/studio` runs and the kit refresh remain. |
+| # | Brief | Verified by |
+| --- | --- | --- |
+| 07 | Final live verification and submit checklist | Every A-E check re-run against live production. Verdict recorded in `openai-submission.md` §0 and the `TRACKER.md` OpenAI row: **not ready**, held on the GCP billing hold (vision QA fails open when no rung answers) plus an unshipped deploy. Four defects found and fixed on the way: the widget fetched its GLB from an origin that refuses CORS to any host but ours, the pending ETA reported a fabricated "5s to go" indefinitely, the vision-QA chain recorded only its first rung's error, and 14 reviewer-visible strings carried a banned em-dash. |
+
+07 was the last brief. Everything still open is the owner's: clear the GCP billing hold
+on project `93741856042`, deploy the tree, delete the duplicate draft GPT (no API for
+it), confirm the portal's screenshot dimensions, and make the partner-portal submit.
-07 is the last brief. The human steps at its end (partner-portal submit, deleting the
-draft GPT) stay with the owner.
+This file stays until nothing under `prompts/finish/` links to it; two `_context/` files
+still do (`00-RETIRED-BY-OWNER.md`, `production-100-00-INDEX.md`), so deleting it here
+would leave two dead links and fail `npm run audit:docs`.
## Rules every task must follow (from `CLAUDE.md`)