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.

fix(smart-money): raise the judge budget so wallet reputation stops losing coins to retention

nirholas671b6094e8a2
diff --git a/api/cron/smart-money-rollup.js b/api/cron/smart-money-rollup.js
index 987a1b402..80e0d6f3e 100644
--- a/api/cron/smart-money-rollup.js
+++ b/api/cron/smart-money-rollup.js
@@ -26,7 +26,18 @@ import { requireCron } from '../_lib/cron-auth.js';
const NETWORK = 'mainnet';
const JUDGE_AFTER_HOURS = 6; // a coin must be this old before "never graduated" = dud
const MAX_AGE_DAYS = 14; // don't reach back further than this
-const COINS_PER_RUN = 80; // judge+fold at most this many coins per run
+// Judge+fold at most this many coins per run. The budget must OUTRUN the
+// firehose or the ground truth silently rots: coin-intel-observe ingests ~32k
+// mints/day, and at the old 80/run (23k/day across 288 runs) the judge fell
+// ~8k/day behind. Because judgeAndFold walks oldest-first and MAX_AGE_DAYS
+// matches the db-retention window, a growing backlog does not queue: the coins
+// at the tail age past 14 days, get pruned, and their buyers never fold into
+// wallet_reputation at all. Measured 2026-08-16: 309k unjudged, oldest already
+// 8.4 days old and climbing toward that cliff. 400/run (115k/day) drains the
+// backlog in a few days and then idles cheaply, since the candidate query
+// returns only unjudged coins. Cost is row volume, not round trips: every
+// phase here is batched into a fixed number of queries whatever the budget.
+const COINS_PER_RUN = 400;
const TOP_WALLETS = 60; // per coin, cap wallets folded/scored
const LIVE_WINDOW_HOURS = 3; // score coins launched within this window
const LIVE_COINS = 150;
diff --git a/prompts/swarm-100/cron-batch-03.md b/prompts/swarm-100/cron-batch-03.md
deleted file mode 100644
index 2f83582c0..000000000
--- a/prompts/swarm-100/cron-batch-03.md
+++ /dev/null
@@ -1,106 +0,0 @@
-# Cron audit batch 03: /api/cron/reconstruct-sweep and 7 more
-
-How to run: paste this file's repo path into a fresh Claude Code chat in this
-repository and say "run this work order". This file is fully self-contained:
-it depends on no other prompt file anywhere. If neighboring files in
-prompts/swarm-100/ are gone, that work is done; if present, ignore them.
-Every claim below rots; step 0 re-measures, and what you measure wins.
-
-## Operating clause (binding)
-
-- Read CLAUDE.md first. Its rules override 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; it never becomes a question that halts work.
-- The only permitted stops are the CLAUDE.md stop-and-ask gates: spending real
- funds or any irreversible on-chain write, git push or a production deploy,
- committing content that references a crypto project other than $THREE, and
- destroying unrecoverable data.
-- No mocks, no fake data, no placeholder stubs, no unfinished-work markers, no
- commented-out code. Real APIs and real integrations only.
-- The em-dash and en-dash characters are banned in everything you write.
-- Concurrent agents share this worktree: stage explicit paths only (never a
- bare add-everything), and commit finished work promptly.
-- Before committing, run: npm run check:rules -- --paths <files you touched>.
- It must exit 0.
-
-## Step 0: re-derive the current state
-
-1. Confirm each cron below still exists in vercel.json crons; removed ones are
- out of scope.
-2. Scheduler truth (crons run on Cloud Scheduler, never in-process):
-
- export PATH="$HOME/google-cloud-sdk/bin:$PATH"
- gcloud scheduler jobs list --project aerial-vehicle-466722-p5 --location us-central1 | grep -E 'reconstruct-sweep|smart-money-rollup|db-retention|smart-money-graph|radar-watchlist|intel-learn|oracle-score|gmgn-seed'
-
-3. Production behavior, last 72h, per cron:
-
- gcloud logging read 'resource.type="cloud_run_revision" resource.labels.service_name="three-ws-api" textPayload:"<cron name>"' --freshness=72h --limit 50 --project aerial-vehicle-466722-p5
-
-## Task: verify and fix these scheduled jobs
-
-Crons in this batch:
-
-- /api/cron/reconstruct-sweep
-- /api/cron/smart-money-rollup
-- /api/cron/db-retention
-- /api/cron/smart-money-graph
-- /api/cron/radar-watchlist
-- /api/cron/intel-learn
-- /api/cron/oracle-score
-- /api/cron/gmgn-seed
-
-For each cron:
-
-1. Read the handler and its auth guard (crons authenticate; find the shared
- pattern in api/_lib before invoking anything).
-2. Invoke it once against a locally started server with the proper cron auth
- from .env. It must complete without an unhandled error and be safe to run
- twice (idempotent); if a local run would move real funds or write on-chain,
- audit its dry-run or validation path instead and say so in the report.
-3. Production logs from step 0: zero unhandled errors and evidence it actually
- fires on schedule. A cron that exists in vercel.json but has no scheduler
- job, or vice versa, is drift: fix it (npm run check:cron-drift helps).
-4. Fix every failure at root: the handler, the schedule, or the env it needs.
-
-## Definition of done
-
-- [ ] Every cron in the batch has a scheduler job, a clean local invocation
- (or an audited dry-run path for money-moving crons), and 72h of
- production logs free of unhandled errors.
-- [ ] npm run check:cron-drift and npm run check:cron-syntax pass.
-- [ ] npm run check:rules -- --paths <touched files> exits 0; vitest passes if
- code changed.
-
-## Never blocked
-
-| Blocker | Resolution (act, do not ask) |
-|---|---|
-| Missing env var or credential | Check .env and .env.local, then the Cloud Run service env, then Secret Manager (the CLAUDE.md self-unblock playbook has the exact commands). If it truly exists nowhere, wire the code fully behind the env var, prove the wiring with a dry run, and name the single missing var in the report. |
-| gcloud not on PATH | Run: export PATH="$HOME/google-cloud-sdk/bin:$PATH" |
-| Dev server port 3000 busy | Another agent may be serving this repo; probe it and reuse if so. Otherwise start your own on a free port: npx vite --port 3101 |
-| Playwright browser missing | npx playwright install chromium |
-| A surface needs a signed-in user | Register a fresh account through the real /register flow against the real API and use it. Never mock the session. |
-| A defect sits in code you did not touch | Fix it if it blocks a Definition of done line (root cause it, never mask it). Otherwise note it in the report and continue. |
-| An unrelated test is red | Same rule. Never pipe npm test through tail; it masks exit codes. |
-
-## 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 (house
- style: type(scope): what changed and why a reader cares).
-3. Delete this prompt file in that same commit:
-
- git rm prompts/swarm-100/cron-batch-03.md
-
- The shrinking directory is this campaign's only progress ledger; there is
- no progress log to update.
-4. Final report: what step 0 measured, what changed (file list), evidence per
- Definition of done line, and any one-line judgment calls. No trailing
- questions, no unexecuted plans.
-
-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.