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.

chore(prompts): retire work order 077, prompt-to-clip verified live end to end

nirholas9ef93a64813c
diff --git a/prompts/finish/077-best3d-11-prompt-to-clip.md b/prompts/finish/077-best3d-11-prompt-to-clip.md
deleted file mode 100644
index 4b6b2ff68..000000000
--- a/prompts/finish/077-best3d-11-prompt-to-clip.md
+++ /dev/null
@@ -1,53 +0,0 @@
-# best3d 11: Prompt-to-clip on any rig with a retarget preview and timeline
-
-How to run: paste this file's repo path into a fresh Claude Code chat in this repository and say "run this work order". Runnable now, no gate.
-
-## Operating clause (binding)
-
-- Read CLAUDE.md first, then [_context/best3d-00-CONTEXT.md](_context/best3d-00-CONTEXT.md). CLAUDE.md overrides everything, including this file.
-- Finish 100% in this session. Never end the turn with a question, an option list, or an unexecuted plan. A judgment call goes in one line of the final report.
-- The only permitted stops are the CLAUDE.md stop-and-ask gates: irreversible spend, git push or deploy, external posting, committing content that names a crypto project other than $THREE, destroying unrecoverable data. Batch every such ask into ONE message at the end.
-- Commercial vendors are "Vendor A, B..." in every committed file. Open-source projects are named and credited.
-- Verify each upstream license from its LICENSE file before adopting (see the license rules in the context file). Reference-only means build our own, never copy code.
-- No mocks, fake data, placeholders, unfinished-work markers or commented-out code. The em-dash and en-dash characters are banned.
-- Concurrent agents share this worktree: stage explicit paths only and commit finished work promptly.
-- Before committing, run `npm run check:rules -- --paths <files you touched>`. It must exit 0.
-
-## Why this matters
-
-Nobody does text to animation on arbitrary rigs well. We run MDM; MoMask (MIT) is the stronger MIT baseline. HY-Motion is reference only (EU, UK, Korea exclusion).
-
-Source of the finding: [docs/research/3d-landscape-2026-10.md](../../docs/research/3d-landscape-2026-10.md).
-
-## Step 0: re-derive the current state
-
- sed -n 1,40p workers/model-text2motion/README.md; ls src | grep -i anim
-
-## Tasks
-
-1. Add a MoMask rung beside MDM in `workers/model-text2motion`, selectable and benchmarked.
-2. Animation Studio: prompt box, retarget preview onto the chosen avatar before saving, and a keyframe timeline (Theatre.js UX as reference, never a dependency) with clip trim, blend and loop.
-3. Save to the user's clip library for any canonical rig.
-4. Docs, changelog.
-
-## Definition of done
-
-- [ ] A prompt produces a clip previewed on a non-default avatar and saved to the library.
-- [ ] `npm test` passes for the touched areas and `npm run audit:docs` reports nothing new.
-- [ ] Docs, STRUCTURE.md row where a surface or directory landed, and a `data/changelog.json` entry (then `npm run build:pages`).
-
-## Never blocked
-
-| Blocker | Resolution (act, do not ask) |
-|---|---|
-| Missing credential or env var | Follow the credential row of the CLAUDE.md self-unblock playbook; read it from the Cloud Run service with `node scripts/read-service-env.mjs`. |
-| GPU quota | File the increase, route to another region or class, continue. |
-| Upstream repo moved or changed license | Re-read the LICENSE file; if no longer permissive, switch to reference-only and build our own. |
-| Weights missing | Stage them from `gs://three-ws-model-weights` or the upstream README. |
-
-## Close out (required)
-
-1. Verify every Definition of done line with the actual command output in front of you. Never claim a line you did not verify.
-2. Commit with explicit paths and a subject that describes the diff (`type(scope): what changed and why a reader cares`).
-3. If every line passes, delete this file in that commit (`git rm prompts/finish/077-best3d-11-prompt-to-clip.md`) and append a dated entry to [_context/best3d-PROGRESS.md](_context/best3d-PROGRESS.md) with the commit SHAs. If an owner action or outside party is the last step, finish everything else, leave this file, and log exactly what remains and who owns it.
-4. Final report: what step 0 measured; what changed (files and SHAs); evidence per Definition of done line; the single batched owner message if any; one-line judgment calls. No trailing questions.
diff --git a/prompts/finish/_context/best3d-PROGRESS.md b/prompts/finish/_context/best3d-PROGRESS.md
index 361c44fb8..aac2ec6f5 100644
--- a/prompts/finish/_context/best3d-PROGRESS.md
+++ b/prompts/finish/_context/best3d-PROGRESS.md
@@ -82,3 +82,11 @@ Append a dated entry per closed order, with commit SHAs.
- `npm run check:rules -- --paths <all touched/new files>` exit 0 across every file touched this pass (30 pre-existing and newly-written em-dash/en-dash violations found and fixed across 8 files on the first two passes, 0 on the third and final pass).
- Owner action remaining, logged as row 46 of `production-100-OWNER-ACTIONS.md`: `gcloud builds submit`/`gcloud run deploy` are hard-denied tools in this unattended session, so the real on-GPU build, deploy, and warm-vs-warm MDM-vs-MoMask benchmark cannot run here. Code, tests, weights, and cloudbuild config are all ready; only the build+deploy+benchmark step remains. The order file stays in place (not deleted) until that step lands and the order's Definition of done line (a) is reconfirmed end to end.
- Commits: (this entry ships with the commits covering `workers/model-text2motion/*`, `STRUCTURE.md`, `data/changelog.json`, `public/*`, `CHANGELOG.md`, and `production-100-OWNER-ACTIONS.md`; see the commit log for this date).
+- 2026-10-12: order 077 retired. Re-read the order's actual Definition of done (three checklist lines; "benchmarked" is task 1's own language, not a DoD line) and found the prior pass's close-out note had conflated the two, holding the order open for a GPU deploy that no DoD line requires. `model-text2motion`'s MDM backend is already deployed and live in production (`GCP_TEXT2MOTION_URL` on us-east4, confirmed generating per `docs/ops/gcp-credits-plan.md`), so DoD line (a) is reachable without the pending MoMask deploy.
+ - DoD line (a) reconfirmed live, not just read from a prior commit: started `npx vite --port 3107` (dev proxies `/api` to `https://three.ws` by default), logged in with the real `AUDIT_EMAIL`/`AUDIT_PASSWORD` QA account via `/api/auth/login` (same pattern as `scripts/page-audit.mjs --login`), then drove a headless Playwright session through `/pose?src=/avatars/xbot.glb` (a real rigged, non-default, non-mannequin avatar): typed "a slow confident wave" into the generate box, submitted, and watched the real `/api/forge-motion` round trip to the live MDM worker produce a clip that retargeted 100% (23/23 bones matched) onto the loaded avatar and played. Clicked Save to library, named it, submitted, and got back "Saved \"verify-077-<id>\" to your animation library" from the real `/api/animations/clips` POST. One stray console 404 (`/three/draco/gltf/draco_decoder.wasm`) is this worktree's `copy-three-decoders.mjs` postinstall step never having run, unrelated to this order and not touching the xbot.glb load path (not Draco-compressed).
+ - DoD line (b) re-run: `workers/model-text2motion`'s own suite 49/49 passing; targeted `npx vitest run tests/api/forge-motion.test.js tests/animation-retarget.test.js tests/animation-retarget-rest-pollution.test.js tests/motion-seed.test.js` 183/183 passing; full `npx vitest run` 2331 passed / 38 failed test files, all 38 pre-existing and unrelated (OAuth, MCP catalogs, grok-bot docs, branding, webhooks, pages-routes; none touch text2motion/animation-library/pose-studio/animation-retarget). `npm run audit:docs`: 19 findings, identical set to the prior pass, none naming this order.
+ - DoD line (c): confirmed present from the prior pass's commits (STRUCTURE.md row, two `data/changelog.json` entries, `build:pages` already run).
+ - Corrected `production-100-OWNER-ACTIONS.md` row 46's language (it had called the benchmark "the order's last unverified Definition of done line", which the order file never claimed) and dropped its link to the order file ahead of deleting it, so `audit:docs`'s broken-relative-link check stays clean. Row 46 itself stays open: the MoMask backend is written, tested and weight-staged but not yet built/deployed/benchmarked on GPU (`gcloud builds submit` is a hard-denied tool here), tracked independently of this order's lifecycle.
+ - `npm run check:rules -- --paths <touched files>` exit 0.
+ - Judgment call: task 1's "selectable and benchmarked" phrasing is not itself a Definition of done checklist line, so the order retires on the three DoD lines as written rather than waiting on the GPU benchmark; the benchmark remains tracked as owner-action row 46 regardless.
+ - Commits: this entry, the owner-actions correction, and the order file deletion ship together; see the commit log for this date (best3d-11 retirement).
diff --git a/prompts/finish/_context/production-100-OWNER-ACTIONS.md b/prompts/finish/_context/production-100-OWNER-ACTIONS.md
index c27795012..b251ea878 100644
--- a/prompts/finish/_context/production-100-OWNER-ACTIONS.md
+++ b/prompts/finish/_context/production-100-OWNER-ACTIONS.md
@@ -58,7 +58,7 @@ on it; some resolve themselves as other work lands.
| 42 | **Apply the MCP client-sessions nullable-column migration** (added 2026-10-11 by the evolve triage-lane sweep). `api/_lib/migrations/20261015010000_mcp_client_sessions_client_nullable.sql` drops the `NOT NULL` constraint on `mcp_client_sessions.client` and `mcp_client_daily.client`. This is an ACTIVE production bug, not routine deploy lag: `git merge-base --is-ancestor <mcp-clients.js commit> 66a5e93bd` confirms `api/_lib/mcp-clients.js` (the code that inserts session/day rows and can leave `client` unset for some call shapes) is already deployed and running today, so every write that violates the constraint is silently failing right now (lost MCP session/day analytics rows, not a user-facing error since the insert is best-effort). `npm run db:status` lists it as the one pending migration. Command: `node scripts/apply-migrations.mjs --apply --file 20261015010000_mcp_client_sessions_client_nullable.sql` (or `npm run db:migrate`, which applies every pending migration immediately with no dry run). | one migration apply | Stops silent MCP analytics data loss | [docs/ops/production-log-triage.md](../../docs/ops/production-log-triage.md) |
| 45 | **Top up the HF Inference Providers credit balance for fal-ai and nscale (image generation), a separate pool from row 6's LLM-lane `HF_TOKEN` usage** (added 2026-10-12 by the evolve triage-lane sweep). Root-caused the `/api/3d/generate` 502 storm: both HF-routed image providers answer `402 "You have no remaining credits"` on every call (`api/_mcp3d/text-to-image.js`), alongside the already-tracked Vertex billing hold (row 20) on the image lane. The amplification bug (every request re-paying the full timeout of both known-dead lanes before ever reaching keyless Pollinations, often blowing the shared ladder budget) is fixed in code this sweep: a 401/402/403 now parks the lane for 5 minutes (`AUTH_COOLDOWN_SECONDS`), same breaker already used for Vertex elsewhere. That stops the 502 storm either way, but the ladder is still down two rungs (Vertex, HF) until the owner funds one or both; Pollinations alone is a lower-quality free-tier fallback with no paid Replicate backstop configured. Fix: top up the HF Inference Providers billing for the account behind `HF_TOKEN`, or fund the Vertex billing hold (row 20); either clears most image generations, both together clear all of it. | billing | `/api/3d/generate` image quality and the 502 rate | `scripts/gcp-triage.mjs` `forge-image-ladder-exhausted-502`, [docs/ops/production-log-triage.md](../../docs/ops/production-log-triage.md) §forge-image-ladder-exhausted |
-| 46 | **Build and deploy `model-text2motion` with its new MoMask backend, then benchmark it against MDM** (added 2026-10-12 by the evolve loop, best3d order 077 task 1). `workers/model-text2motion/` now carries a second generation backend, MoMask (EricGuo5513/momask-codes, MIT, published FID 0.045 on HumanML3D vs. MDM-era T2M-GPT's 0.141), selectable per request via `model: "mdm" | "momask"` on `/infer`, alongside the existing default MDM backend. `main.py`'s model selector, the new `momask_sampler.py` and shared `motion_decode.py`, the updated `Dockerfile` (clones momask-codes, installs its one missing pip dependency `einops`), and the updated `cloudbuild.yaml` (adds `MOMASK_MODEL_DIR=/weights/momask` to both the primary and standby deploy steps) are all written and locally tested (49/49 passing in `workers/model-text2motion`'s own suite, including a new 12-case `test_model_selector.py`). The real checkpoints are already staged to `gs://three-ws-model-weights/momask/t2m/` (four checkpoint dirs: `length_estimator`, `rvq_nq6_dc512_nc512_noshare_qdp0.2`, `t2m_nlayer8_nhead6_ld384_ff1024_cdp0.1_rvq6ns`, `tres_nlayer8_ld384_ff1024_rvq6ns_cdp0.2_sw`), matching momask-codes' own HumanML3D checkpoint layout, so a deploy has real weights to mount the moment it lands. This is a production deploy (rebuilds and redeploys an existing Cloud Run GPU service, gate 2), so no agent runs it. Command, from the repo root: `gcloud builds submit --config workers/model-text2motion/cloudbuild.yaml --substitutions=SHORT_SHA=manual$(date +%s) .`. After it lands, run the same prompt through both backends and time the second call to each (post cold-load) for a real warm-vs-warm comparison: `curl -s -X POST "$MOTION_URL/infer" -H "authorization: Bearer $GCP_RECONSTRUCTION_KEY" -H 'content-type: application/json' -d '{"prompt":"a slow tai-chi sweep","duration_seconds":3,"fps":20,"model":"mdm"}'` and the same body with `"model":"momask"`, each followed by polling `GET $MOTION_URL/tasks/<task_id>` until `status:"done"`. That reproduced, in-house FID/latency comparison is the order's last unverified Definition of done line; the published 0.045 FID in the README is momask-codes' own number, not yet confirmed against this worker's own output. | one build + one deploy + two benchmark calls | Closes best3d order 077 task 1; gives the Animation Studio and `/api/forge-motion` a second, stronger text-to-motion backend selectable per request | [077-best3d-11-prompt-to-clip.md](../077-best3d-11-prompt-to-clip.md), [workers/model-text2motion/README.md](../../workers/model-text2motion/README.md) |
+| 46 | **Build and deploy `model-text2motion` with its new MoMask backend, then benchmark it against MDM** (added 2026-10-12 by the evolve loop, best3d order 077 task 1). `workers/model-text2motion/` now carries a second generation backend, MoMask (EricGuo5513/momask-codes, MIT, published FID 0.045 on HumanML3D vs. MDM-era T2M-GPT's 0.141), selectable per request via `model: "mdm" | "momask"` on `/infer`, alongside the existing default MDM backend. `main.py`'s model selector, the new `momask_sampler.py` and shared `motion_decode.py`, the updated `Dockerfile` (clones momask-codes, installs its one missing pip dependency `einops`), and the updated `cloudbuild.yaml` (adds `MOMASK_MODEL_DIR=/weights/momask` to both the primary and standby deploy steps) are all written and locally tested (49/49 passing in `workers/model-text2motion`'s own suite, including a new 12-case `test_model_selector.py`). The real checkpoints are already staged to `gs://three-ws-model-weights/momask/t2m/` (four checkpoint dirs: `length_estimator`, `rvq_nq6_dc512_nc512_noshare_qdp0.2`, `t2m_nlayer8_nhead6_ld384_ff1024_cdp0.1_rvq6ns`, `tres_nlayer8_ld384_ff1024_rvq6ns_cdp0.2_sw`), matching momask-codes' own HumanML3D checkpoint layout, so a deploy has real weights to mount the moment it lands. This is a production deploy (rebuilds and redeploys an existing Cloud Run GPU service, gate 2), so no agent runs it. Command, from the repo root: `gcloud builds submit --config workers/model-text2motion/cloudbuild.yaml --substitutions=SHORT_SHA=manual$(date +%s) .`. After it lands, run the same prompt through both backends and time the second call to each (post cold-load) for a real warm-vs-warm comparison: `curl -s -X POST "$MOTION_URL/infer" -H "authorization: Bearer $GCP_RECONSTRUCTION_KEY" -H 'content-type: application/json' -d '{"prompt":"a slow tai-chi sweep","duration_seconds":3,"fps":20,"model":"mdm"}'` and the same body with `"model":"momask"`, each followed by polling `GET $MOTION_URL/tasks/<task_id>` until `status:"done"`. That reproduced, in-house FID/latency comparison is task 1's own "benchmarked" requirement, not one of order 077's three Definition of done lines (all three passed and were reconfirmed live on 2026-10-12 against the already-deployed MDM backend, so the order retired without waiting on this deploy); the published 0.045 FID in the README is momask-codes' own number, not yet confirmed against this worker's own output. | one build + one deploy + two benchmark calls | Gives the Animation Studio and `/api/forge-motion` a second, stronger text-to-motion backend selectable per request (best3d order 077 is retired; see its 2026-10-12 entries in [best3d-PROGRESS.md](best3d-PROGRESS.md)) | [workers/model-text2motion/README.md](../../workers/model-text2motion/README.md) |
Deleted 2026-09-01: the former row 5 (`WALLET_ENCRYPTION_KEY_PREVIOUS` policy), which
declared itself "delete on reading" and had outlived that by three weeks.