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.

style(packages): purge banned dashes and fix README accuracy across the batch-08 SDK packages

nirholasccae31ad500f
diff --git a/packages/pumpfun-skills/README.md b/packages/pumpfun-skills/README.md
index 5938b7218..a89573a04 100644
--- a/packages/pumpfun-skills/README.md
+++ b/packages/pumpfun-skills/README.md
@@ -4,7 +4,7 @@
<h1 align="center">@three-ws/pumpfun-skills</h1>
-<p align="center"><strong>pump.fun launch + trade as composable agent tools — create a coin, swap, and read creator fees, with a runtime-supplied mint.</strong></p>
+<p align="center"><strong>pump.fun launch + trade as composable agent tools: create a coin, swap, and read creator fees, with a runtime-supplied mint.</strong></p>
<p align="center">
<a href="https://www.npmjs.com/package/@three-ws/pumpfun-skills"><img alt="npm" src="https://img.shields.io/npm/v/@three-ws/pumpfun-skills?logo=npm&color=cb3837"></a>
@@ -24,13 +24,13 @@
---
> `@three-ws/pumpfun-skills` is the official client for the three.ws **pump.fun
-> skill set** — the same launch/trade/fee plumbing the three.ws agents use,
+> skill set**: the same launch/trade/fee plumbing the three.ws agents use,
> exposed as plain functions instead of [Agent Skills](https://agentskills.io)
> tools. It wraps pump.fun's agent transaction API
> (`https://fun-block.pump.fun/agents/*`), which builds the unsigned or
> mint-signed Solana transaction for you, and the public coins read API
> (`frontend-api-v3.pump.fun/coins-v2`). Every function is **coin-agnostic**:
-> the mint, amounts, and wallet are supplied at call time — nothing is
+> the mint, amounts, and wallet are supplied at call time; nothing is
> hardcoded. You build, you co-sign, you broadcast. It pairs with
> [`@three-ws/agent-payments`](https://www.npmjs.com/package/@three-ws/agent-payments)
> for tokenized-agent invoices and the three.ws launch feed.
@@ -40,11 +40,11 @@
Launching and trading on pump.fun by hand means juggling two SDKs
(`@pump-fun/pump-sdk`, `@pump-fun/pump-swap-sdk`), deriving PDAs, detecting
whether a coin is still on the bonding curve or has graduated to an AMM pool,
-quoting slippage, and assembling a versioned transaction — before you ever ask
+quoting slippage, and assembling a versioned transaction, all before you ever ask
the user to sign. These skills collapse that into three calls:
- **One call, a ready transaction.** `createCoin({...})` returns a
- base64 transaction with the mint keypair already signed — you only co-sign
+ base64 transaction with the mint keypair already signed; you only co-sign
with the wallet and submit.
- **State-aware swaps.** `swap({...})` auto-detects bonding curve vs. graduated
AMM and builds the correct route, with slippage and optional Jito MEV
@@ -53,10 +53,10 @@ the user to sign. These skills collapse that into three calls:
(creator, cashback, or a sharing config), reports vault balances, and
`collectFees({...})` builds the claim or distribution transaction.
- **Mint supplied at runtime.** Pass any pump.fun mint you control. There is no
- coin list, no allowlist — generic plumbing, your inputs.
+ coin list, no allowlist: generic plumbing, your inputs.
This is the SDK twin of the `pumpfun_create_coin`, `pumpfun_swap`, and
-`pumpfun_collect_fees` MCP/skill tools — same endpoints, plain functions.
+`pumpfun_collect_fees` MCP/skill tools: same endpoints, plain functions.
## Install
@@ -67,7 +67,7 @@ npm install @three-ws/pumpfun-skills
Works in Node 18+ and the browser (uses `fetch`). To sign and broadcast the
transactions these functions return, add `@solana/web3.js`. The fee-reading
calls go through pump.fun's read API; the build calls go through pump.fun's
-agent API — no key, no three.ws account required.
+agent API; no key, no three.ws account required.
## Quick start
@@ -190,7 +190,7 @@ bonding-curve, pool, cashback, and sharing-config state. No transaction, no RPC.
### `collectFees(input) → Promise<BuiltTx>`
Build a transaction to collect creator fees, or distribute them via the sharing
-config — auto-detected from on-chain state. Wraps `POST /agents/collect-fees`.
+config, auto-detected from on-chain state. Wraps `POST /agents/collect-fees`.
| Field | Type | Notes |
|---|---|---|
@@ -216,7 +216,7 @@ exactly `10000`). Mode is auto-detected if omitted. Wraps
## How it works
-Two surfaces, one rule — **the function builds, you sign**:
+Two surfaces, one rule (**the function builds, you sign**):
```
createCoin / swap coinFees collectFees / sharingConfig
@@ -238,7 +238,7 @@ base64 transaction FeeInfo JSON base64 transaction
- **Build calls** post your inputs (plus `encoding: 'base64'`) to the pump.fun
agent API and get back a transaction. `create-coin` returns it with the new
- **mint keypair already signed** — you only add the wallet signature. `swap`,
+ **mint keypair already signed**: you only add the wallet signature. `swap`,
`collect-fees`, and `sharing-config` return a transaction for the wallet to
sign outright.
- **Read calls** (`coinFees`) hit `coins-v2`, which carries the bonding-curve,
@@ -246,7 +246,7 @@ base64 transaction FeeInfo JSON base64 transaction
balances from it. No wallet, no signing, no RPC.
- **State detection is automatic.** `swap` and `collectFees` inspect whether the
coin is still on the bonding curve or graduated to an AMM pool and build the
- correct route — you never branch on it yourself.
+ correct route; you never branch on it yourself.
Override the read backend (e.g. for devnet) with the `PUMP_COINS_V2_BASE`
environment variable; the build API base is `https://fun-block.pump.fun/agents`.
@@ -267,12 +267,12 @@ never swallowed:
| `coinFees` | empty `coins-v2` body | Wrong cluster, or coin missing. | Set `PUMP_COINS_V2_BASE` for devnet. |
| any | `frontRunningProtection: true`, no `tipAmount` | Jito route needs a tip. | Pass `tipAmount` (SOL). |
-A built transaction is **partially signed at most** — it is never broadcast for
+A built transaction is **partially signed at most**: it is never broadcast for
you. Submitting (and paying network fees) is always the caller's step.
## Examples
-**Read before you claim** — never build a no-op collection:
+**Read before you claim**: never build a no-op collection:
```js
import { coinFees, collectFees } from '@three-ws/pumpfun-skills';
@@ -318,9 +318,9 @@ const { transaction } = await swap({
## Related
-- [`@three-ws/agent-payments`](https://www.npmjs.com/package/@three-ws/agent-payments) — tokenized-agent invoices and on-chain payment verification.
-- [`@three-ws/x402-fetch`](https://www.npmjs.com/package/@three-ws/x402-fetch) — pay-per-call USDC settlement for agent services.
-- [`@three-ws/forge`](https://www.npmjs.com/package/@three-ws/forge) — generate the 3D avatar that fronts a tokenized agent.
+- [`@three-ws/agent-payments`](https://www.npmjs.com/package/@three-ws/agent-payments): tokenized-agent invoices and on-chain payment verification.
+- [`@three-ws/x402-fetch`](https://www.npmjs.com/package/@three-ws/x402-fetch): pay-per-call USDC settlement for agent services.
+- [`@three-ws/forge`](https://www.npmjs.com/package/@three-ws/forge): generate the 3D avatar that fronts a tokenized agent.
---
diff --git a/packages/pumpfun-skills/package.json b/packages/pumpfun-skills/package.json
index 465f3de42..f4d354353 100644
--- a/packages/pumpfun-skills/package.json
+++ b/packages/pumpfun-skills/package.json
@@ -1,7 +1,7 @@
{
"name": "@three-ws/pumpfun-skills",
"version": "0.1.1",
- "description": "pump.fun launch + trade skills as composable agent tools. Create a coin, swap on the bonding curve or AMM, and read creator fees — with a runtime-supplied mint.",
+ "description": "pump.fun launch + trade skills as composable agent tools. Create a coin, swap on the bonding curve or AMM, and read creator fees, with a runtime-supplied mint.",
"type": "module",
"main": "./src/index.js",
"module": "./src/index.js",
diff --git a/packages/pumpfun-skills/src/http.js b/packages/pumpfun-skills/src/http.js
index 6398a3e53..339c074eb 100644
--- a/packages/pumpfun-skills/src/http.js
+++ b/packages/pumpfun-skills/src/http.js
@@ -20,7 +20,7 @@ export class ThreeWsError extends Error {
/**
* Thrown on HTTP 402. The endpoint wants payment before it will do the work.
- * `accepts` is the x402 challenge (asset/amount/network/payTo) when present —
+ * `accepts` is the x402 challenge (asset/amount/network/payTo) when present;
* pass a payment-aware `fetch` (e.g. @three-ws/x402-fetch) to settle it
* automatically, or read `accepts` and pay it yourself.
*/
@@ -63,7 +63,7 @@ export function createHttp(opts = {}) {
const baseUrl = resolveBaseUrl(opts.baseUrl);
const fetchImpl = opts.fetch || (typeof globalThis !== 'undefined' ? globalThis.fetch : undefined);
if (typeof fetchImpl !== 'function') {
- throw new ThreeWsError('No fetch implementation available — run on Node 18+ or pass { fetch }.', { code: 'no_fetch' });
+ throw new ThreeWsError('No fetch implementation available: run on Node 18+ or pass { fetch }.', { code: 'no_fetch' });
}
const baseHeaders = { accept: 'application/json', ...(opts.headers || {}) };
if (opts.apiKey) baseHeaders.authorization = `Bearer ${opts.apiKey}`;
diff --git a/packages/pumpfun-skills/src/index.js b/packages/pumpfun-skills/src/index.js
index fbb0f72d9..92ebc0d21 100644
--- a/packages/pumpfun-skills/src/index.js
+++ b/packages/pumpfun-skills/src/index.js
@@ -1,15 +1,15 @@
-// @three-ws/pumpfun-skills — pump.fun launch + trade as composable agent tools.
+// @three-ws/pumpfun-skills: pump.fun launch + trade as composable agent tools.
// The SDK twin of the pumpfun_create_coin / pumpfun_swap / pumpfun_collect_fees
// skills: every build call posts your inputs to pump.fun's agent transaction API
// and hands back a base64 transaction you co-sign and broadcast; coinFees reads
// the public coins-v2 API. The mint, amounts, and wallet are supplied at call
-// time — nothing is hardcoded. See README.md for the full reference.
+// time, nothing is hardcoded. See README.md for the full reference.
import { createHttp, ThreeWsError } from './http.js';
export { ThreeWsError, PaymentRequiredError, DEFAULT_BASE_URL } from './http.js';
-/** pump.fun's agent transaction API — builds the unsigned/mint-signed tx. */
+/** pump.fun's agent transaction API: builds the unsigned/mint-signed tx. */
export const AGENT_API_BASE = 'https://fun-block.pump.fun/agents';
/** pump.fun's public coins read API. Override with PUMP_COINS_V2_BASE. */
export const COINS_V2_BASE = 'https://frontend-api-v3.pump.fun/coins-v2';
@@ -22,8 +22,8 @@ const FEE_DESTINATIONS = ['creator', 'cashback', 'sharing_config'];
* Create a pump.fun skills client bound to a transport, agent API base, and
* coins read base. For most callers the default exports
* (`createCoin` / `swap` / `coinFees` / `collectFees` / `sharingConfig`) are
- * enough; use this to reuse configuration — a payment-aware `fetch`, a devnet
- * read backend, or a custom agent origin — across many calls.
+ * enough; use this to reuse configuration (a payment-aware `fetch`, a devnet
+ * read backend, or a custom agent origin) across many calls.
*
* @param {object} [options]
* @param {string} [options.agentBaseUrl] Agent build API origin (default fun-block.pump.fun/agents).
@@ -289,7 +289,7 @@ function validateShareholders(shareholders) {
}
}
-// Jito routing needs a tip — fail fast before the build call rather than letting
+// Jito routing needs a tip: fail fast before the build call rather than letting
// pump.fun reject it (mirrors the README's edge-case table).
function assertJitoTip(input, fn) {
if (input.frontRunningProtection && (input.tipAmount === undefined || input.tipAmount === null)) {
diff --git a/packages/pumpfun-skills/test/pumpfun-skills.test.js b/packages/pumpfun-skills/test/pumpfun-skills.test.js
index 6af94e8a5..b4471f974 100644
--- a/packages/pumpfun-skills/test/pumpfun-skills.test.js
+++ b/packages/pumpfun-skills/test/pumpfun-skills.test.js
@@ -8,13 +8,13 @@ import {
} from '../src/index.js';
// $THREE is the only coin. Tests use the real $THREE mint or a clearly-synthetic
-// placeholder — never a third-party mint.
+// placeholder, never a third-party mint.
const THREE_MINT = 'FeMbDoX7R1Psc4GEcvJdsbNbZA3bfztcyDCatJVJpump';
const SYNTH_MINT = 'THREEsynthetic1111111111111111111111111111';
const WALLET = 'YourWa11et1111111111111111111111111111111111';
// A scripted fetch double: each call shifts the next queued response and records
-// the request. No network, no real endpoints — we assert on request shaping and
+// the request. No network, no real endpoints, we assert on request shaping and
// response parsing, which is all the SDK is responsible for.
function stubFetch(responses) {
const calls = [];
diff --git a/packages/react/README.md b/packages/react/README.md
index b2499bd4f..d22b4bb33 100644
--- a/packages/react/README.md
+++ b/packages/react/README.md
@@ -1,7 +1,7 @@
# @three-ws/react
React components for embedding [three.ws](https://three.ws) 3D AI agents. Drop a
-walking 3D avatar into any React app in two lines — no Three.js, no WebGL setup,
+walking 3D avatar into any React app in two lines: no Three.js, no WebGL setup,
no build config. The 3D runtime lives inside three.ws and renders through a
sandboxed iframe.
@@ -53,7 +53,7 @@ export function Avatar() {
// Later, drive the live embed:
ref.current.setMotion('walk'); // 'idle' | 'walk' | 'run'
ref.current.setAvatar('av-1234'); // swap avatar live
-ref.current.setSpeed(1.5); // walk-speed multiplier (0.3–3)
+ref.current.setSpeed(1.5); // walk-speed multiplier (0.3-3)
ref.current.narrate('Hi there!'); // speech bubble over the avatar
ref.current.setEnvironment('studio');
ref.current.resetPose();
@@ -63,8 +63,8 @@ ref.current.resetPose();
| Prop | Type | Default | Description |
|---------------|---------------------------------------|----------------|--------------------------------------------------------|
-| `agentId` | `string` (required) | — | The three.ws agent ID — its avatar is rendered. |
-| `avatarId` | `string` | — | Override the agent's default avatar. |
+| `agentId` | `string` (required) | n/a | The three.ws agent ID whose avatar is rendered. |
+| `avatarId` | `string` | n/a | Override the agent's default avatar. |
| `baseUrl` | `string` | `"https://three.ws"` | Embed origin (self-hosted / staging players). |
| `controls` | `"joystick" \| "keyboard" \| "none"` | `"joystick"` | Movement controls. |
| `background` | `string` | transparent | `"transparent"` or a hex color like `"#1b1b1b"`. |
@@ -72,13 +72,13 @@ ref.current.resetPose();
| `autoplay` | `boolean` | `false` | Autoplay an idle walk loop. |
| `ground` | `boolean` | `true` | Show the shadow ground disc. |
| `orbit` | `boolean` | `true` | Allow orbit drag on desktop. |
-| `speed` | `number` | `1` | Walk-speed multiplier (0.3–3), applied once ready. |
+| `speed` | `number` | `1` | Walk-speed multiplier (0.3-3), applied once ready. |
| `width` | `string \| number` | `"100%"` | Container width (number → px). |
| `height` | `string \| number` | `"600px"` | Container height (number → px). |
-| `onLoad` | `() => void` | — | Fires when the 3D scene is ready. |
-| `onError` | `(err: Error) => void` | — | Fires on load failure. |
-| `className` | `string` | — | CSS class on the wrapper `<div>`. |
-| `style` | `React.CSSProperties` | — | Inline styles on the wrapper `<div>`. |
+| `onLoad` | `() => void` | n/a | Fires when the 3D scene is ready. |
+| `onError` | `(err: Error) => void` | n/a | Fires on load failure. |
+| `className` | `string` | n/a | CSS class on the wrapper `<div>`. |
+| `style` | `React.CSSProperties` | n/a | Inline styles on the wrapper `<div>`. |
| `title` | `string` | three.ws title | Accessible title for the underlying iframe. |
### Imperative handle (via `ref`)
@@ -89,7 +89,7 @@ ref.current.resetPose();
| `setAvatar(id)` | Swap the rendered avatar live. |
| `setMotion(motion)` | `"idle" \| "walk" \| "run"`. |
| `setEnvironment(env)` | Switch the environment preset live. |
-| `setSpeed(value)` | Set the walk-speed multiplier (0.3–3). |
+| `setSpeed(value)` | Set the walk-speed multiplier (0.3-3). |
| `narrate(text)` | Show a speech bubble above the avatar. |
| `resetPose()` | Recenter the avatar on the ground. |
| `iframe` | The underlying `HTMLIFrameElement` (or null).|
@@ -97,7 +97,7 @@ ref.current.resetPose();
## Security
Inbound `postMessage` events are accepted **only** from the three.ws origin and
-**only** from this component's own iframe — messages from any other origin or
+**only** from this component's own iframe; messages from any other origin or
window are ignored. This is enforced regardless of which `onLoad`/`onError`
handlers you pass.
@@ -108,7 +108,7 @@ No `@types` install needed.
## Requirements
-`react` and `react-dom` `>= 17` are peer dependencies — the package never bundles
+`react` and `react-dom` `>= 17` are peer dependencies; the package never bundles
its own React.
## Links
diff --git a/packages/react/src/Agent3D.jsx b/packages/react/src/Agent3D.jsx
index 23630b3f4..1bdac861e 100644
--- a/packages/react/src/Agent3D.jsx
+++ b/packages/react/src/Agent3D.jsx
@@ -48,7 +48,7 @@ function buildEmbedSrc({
}
/**
- * <Agent3D> — embeds a three.ws walking 3D AI agent via a sandboxed iframe.
+ * <Agent3D> embeds a three.ws walking 3D AI agent via a sandboxed iframe.
*
* The 3D runtime (Three.js / WebGL) lives inside three.ws; this component is a
* thin, dependency-free wrapper around the embed player and its postMessage
@@ -95,7 +95,7 @@ const Agent3D = forwardRef(function Agent3D(
[baseUrl, agentId, avatarId, controls, background, environment, autoplay, ground, orbit],
);
- // The only origin we trust for inbound messages — derived from the embed URL.
+ // The only origin we trust for inbound messages, derived from the embed URL.
const targetOrigin = useMemo(() => new URL(src).origin, [src]);
const post = useCallback(
diff --git a/packages/react/src/index.d.ts b/packages/react/src/index.d.ts
index 67bb44a1f..3e9e53ffd 100644
--- a/packages/react/src/index.d.ts
+++ b/packages/react/src/index.d.ts
@@ -19,7 +19,7 @@ export interface Agent3DProps {
ground?: boolean;
/** Allow orbit drag on desktop. Defaults to `true`. */
orbit?: boolean;
- /** Walk-speed multiplier (0.3–3). Applied live once the scene is ready. */
+ /** Walk-speed multiplier (0.3-3). Applied live once the scene is ready. */
speed?: number;
/** Container width. Number is treated as px. Defaults to `"100%"`. */
width?: string | number;
@@ -49,7 +49,7 @@ export interface Agent3DHandle {
setMotion: (motion: 'idle' | 'walk' | 'run') => void;
/** Switch the environment preset live. */
setEnvironment: (env: string) => void;
- /** Set the walk-speed multiplier live (0.3–3). */
+ /** Set the walk-speed multiplier live (0.3-3). */
setSpeed: (value: number) => void;
/** Show a speech bubble above the avatar. */
narrate: (text: string) => void;
@@ -63,5 +63,5 @@ export declare const Agent3D: React.ForwardRefExoticComponent<
Agent3DProps & React.RefAttributes<Agent3DHandle>
>;
-/** Alias for {@link Agent3D} — matches existing three.ws embed snippets. */
+/** Alias for {@link Agent3D}, matches existing three.ws embed snippets. */
export declare const WalkEmbed: typeof Agent3D;
diff --git a/packages/react/src/index.js b/packages/react/src/index.js
index a4296b854..440ef1045 100644
--- a/packages/react/src/index.js
+++ b/packages/react/src/index.js
@@ -1,2 +1,2 @@
export { Agent3D } from './Agent3D.jsx';
-export { Agent3D as WalkEmbed } from './Agent3D.jsx'; // alias — matches existing embed snippets
+export { Agent3D as WalkEmbed } from './Agent3D.jsx'; // alias, matches existing embed snippets
diff --git a/packages/react/test/helpers.mjs b/packages/react/test/helpers.mjs
index e8ca43475..6f05a5747 100644
--- a/packages/react/test/helpers.mjs
+++ b/packages/react/test/helpers.mjs
@@ -1,7 +1,7 @@
// Shared helpers for the @three-ws/react test suite.
//
// Tests run against the built bundles in dist/ (the artifact that actually
-// ships) — `npm test` triggers a fresh build via the pretest hook. Rendering
+// ships): `npm test` triggers a fresh build via the pretest hook. Rendering
// is done with react-dom/server so no DOM shim is needed: everything the
// component puts in markup (iframe src, wrapper styles, pass-through props)
// is asserted on the serialized HTML string.
diff --git a/packages/reputation/README.md b/packages/reputation/README.md
index ecdef992a..aac303760 100644
--- a/packages/reputation/README.md
+++ b/packages/reputation/README.md
@@ -4,7 +4,7 @@
<h1 align="center">@three-ws/reputation</h1>
-<p align="center"><strong>Read an agent's trust score, rank the leaderboard, and record on-chain validations — one zero-dependency import over the live three.ws reputation API.</strong></p>
+<p align="center"><strong>Read an agent's trust score, rank the leaderboard, and record on-chain validations, one zero-dependency import over the live three.ws reputation API.</strong></p>
<p align="center">
<a href="https://www.npmjs.com/package/@three-ws/reputation"><img alt="npm" src="https://img.shields.io/npm/v/@three-ws/reputation?logo=npm&color=cb3837"></a>
@@ -26,31 +26,31 @@
> `@three-ws/reputation` is the official client for agent reputation as wired into
> [three.ws](https://three.ws). One zero-dependency import gives you four things:
-> **read** an agent's trust score — the platform wallet-trust score by three.ws
+> **read** an agent's trust score: the platform wallet-trust score by three.ws
> agent UUID, or the on-chain attestation aggregate (feedback, stake, validations,
> disputes) by Solana asset address; **rank** the live leaderboard of trusted
> agents; **read** an agent's latest [ERC-8004](https://eips.ethereum.org/EIPS/eip-8004)
-> validation attestation on any supported EVM chain; and **attest** — run an
+> validation attestation on any supported EVM chain; and **attest**: run an
> agent's GLB through the platform validator and record a signed on-chain
-> validation. Every call is a thin wrapper over a live three.ws endpoint — the same
-> data the trust badges and the `agent_reputation` MCP tool surface — so the score
+> validation. Every call is a thin wrapper over a live three.ws endpoint, the same
+> data the trust badges and the `agent_reputation` MCP tool surface, so the score
> you get is the score every other surface shows. Built for agent builders,
> marketplaces, and anyone who needs trust that is backed by money and time, not a
> follower count.
## Why
-Agent reputation only means something if it is non-gameable and auditable — backed
+Agent reputation only means something if it is non-gameable and auditable, backed
by real on-chain attestations, stake, validations, disputes, and activity, not a
self-declared rating. three.ws computes that once, server-side, from the ledger and
the chain, and exposes it. Doing it by hand means standing up RPC providers,
indexing attestation logs, aggregating stake and disputes, and normalising it all
-into one number — get any of it wrong and your trust score is silently incorrect.
+into one number; get any of it wrong and your trust score is silently incorrect.
This SDK is the one-line front door:
- **One call, a real score.** `reputation(agent)` returns the aggregate score,
- tier, feedback, and stake — read live from the platform, no indexer, no cache,
+ tier, feedback, and stake, read live from the platform: no indexer, no cache,
no fabricated fallback.
- **Read by UUID or asset.** Pass a three.ws agent **UUID** for the platform
wallet-trust score, or a **Solana asset address** for the on-chain attestation
@@ -61,7 +61,7 @@ This SDK is the one-line front door:
platform validator and records a signed on-chain validation, so an agent can
build a verifiable track record.
-This is the SDK twin of the [`agent_reputation` MCP tool](https://three.ws/mcp) —
+This is the SDK twin of the [`agent_reputation` MCP tool](https://three.ws/mcp):
the same data, exposed as plain functions instead of an MCP call.
## Install
@@ -77,7 +77,7 @@ an API token with the `avatars:write` scope.
## Quick start
-Read an agent's trust score by three.ws agent UUID — no key, no wallet:
+Read an agent's trust score by three.ws agent UUID, no key, no wallet:
```js
import { reputation } from '@three-ws/reputation';
@@ -135,10 +135,10 @@ from `GET /api/agents/solana-reputation`). Any other value throws `invalid_input
| Option | Type | Default | Notes |
|---|---|---|---|
-| `network` | `'mainnet' \| 'devnet'` | `'mainnet'` | Solana cluster — only applies to asset-address reads. |
-| `signal` | `AbortSignal` | — | Cancel the in-flight read. |
+| `network` | `'mainnet' \| 'devnet'` | `'mainnet'` | Solana cluster, only applies to asset-address reads. |
+| `signal` | `AbortSignal` | n/a | Cancel the in-flight read. |
-**Returns** — a `WalletReputation` (UUID) or a `SolanaReputation` (asset), each
+**Returns**: a `WalletReputation` (UUID) or a `SolanaReputation` (asset), each
carrying a `.raw` escape hatch to the untouched endpoint JSON. Tell them apart by
`.kind`.
@@ -165,7 +165,7 @@ carrying a `.raw` escape hatch to the untouched endpoint JSON. Tell them apart b
| `agent` | `string \| null` | The asset address read. |
| `network` | `string \| null` | Cluster the read came from. |
| `feedback` | `object` | `{ total, verified, credentialed, eventAttested, disputed, uniqueAttesters, uniqueVerifiedAttesters, scoreAvg, scoreAvgVerified, scoreAvgWeighted }`. |
-| `stake` | `object` | `{ totalLamports, count, uniqueStakers, topStakers[] }`. |
+| `stake` | `object` | `{ totalLamports, count, uniqueStakers, grossLamports, retiredLamports, retiredCount, topStakers[] }`. `totalLamports` is net conviction: settled positions are retired out of it, and the gross/retired fields keep the history readable. |
| `validation` / `tasks` | `object \| null` | Validation + task rollups. |
| `disputesFiled` / `revokedCount` | `number` | Dispute + revocation counts. |
| `tokenActivity` / `pumpPayments` | `object \| null` | On-chain activity signals. |
@@ -174,13 +174,13 @@ carrying a `.raw` escape hatch to the untouched endpoint JSON. Tell them apart b
### `leaderboard(options?) → Promise<Leaderboard>`
Fetch the platform's live ranking of trusted agents
-(`GET /api/reputation/leaderboard`) — every rank is the same non-gameable
+(`GET /api/reputation/leaderboard`): every rank is the same non-gameable
wallet-trust score the badge shows, computed from real ledger + chain activity.
-**Options** — `{ limit?: number, signal?: AbortSignal }`. `limit` is clamped to
-1–50 (default 20).
+**Options**: `{ limit?: number, signal?: AbortSignal }`. `limit` is clamped to
+1-50 (default 20).
-**Returns** — `{ generatedAt, count, scored, agents }`, where each agent carries
+**Returns**: `{ generatedAt, count, scored, agents }`, where each agent carries
`rank`, `id`, `name`, `avatarThumbnailUrl`, `solanaAddress`, `score`, `tier`,
`tierLabel`, `totals`, `agentUrl`, and a `breakdownUrl` linking to the auditable
breakdown.
@@ -188,11 +188,11 @@ breakdown.
### `validation(chainId, agentId, options?) → Promise<ValidationRead>`
Read an agent's latest ERC-8004 validation attestation
-(`GET /api/erc8004/validation`) — the walletless read that powers the "Validated"
+(`GET /api/erc8004/validation`): the walletless read that powers the "Validated"
badge. `chainId` is a chain name or numeric id from [Chains](#chains) (or pass
`options.chain` to override it); `agentId` is the uint ERC-8004 agent id.
-**Returns** `ValidationRead` — `{ chain, chainId, agentId, kind, registry,
+**Returns** `ValidationRead`: `{ chain, chainId, agentId, kind, registry,
available, exists, passed, proofHash, proofURI, proofUrlResolved, validator,
validatorExplorer, validatedAt, reason, raw }`. When no attestation exists,
`exists` is `false` and `passed` is `null`.
@@ -219,26 +219,26 @@ Requires a signed-in account or an `avatars:write`-scoped token (pass it as
| `glbUrl` | `string` | Explicit GLB to validate; resolved from the agent when omitted. |
| `signal` | `AbortSignal` | Cancel the write. |
-**Returns** `AttestReceipt` — `{ lane, status, ok, passed, kind, signature,
-txExplorer, proofHash, proofURI, validator, … , raw }`, where `status` is
+**Returns** `AttestReceipt`: `{ lane, status, ok, passed, kind, signature,
+txExplorer, proofHash, proofURI, validator, …, raw }`, where `status` is
`'minted'` (a new on-chain tx) or `'deduped'` (a prior attestation already landed).
### `createReputation(options?) → Client`
-Build a client bound to a base URL, fetch, and optional auth — reuse it across
+Build a client bound to a base URL, fetch, and optional auth; reuse it across
calls instead of relying on the module-level default exports.
| Option | Type | Default | Notes |
|---|---|---|---|
| `baseUrl` | `string` | `https://three.ws` | Override for self-hosted / preview origins. |
| `fetch` | `typeof fetch` | global `fetch` | Inject a custom (e.g. payment-aware) fetch. |
-| `apiKey` | `string` | — | Bearer token (needs `avatars:write`) for `attest()`. |
-| `headers` | `Record<string,string>` | — | Extra headers on every request. |
+| `apiKey` | `string` | n/a | Bearer token (needs `avatars:write`) for `attest()`. |
+| `headers` | `Record<string,string>` | n/a | Extra headers on every request. |
Also exported: `SUPPORTED_CHAINS` (the frozen chain list), `DEFAULT_BASE_URL`,
and the error classes `ThreeWsError` / `PaymentRequiredError`.
-### Under the hood — raw HTTP
+### Under the hood: raw HTTP
Every function is a thin wrapper over a public three.ws endpoint, so the docs hold
even before the SDK ships in your stack:
@@ -269,14 +269,14 @@ normalises the snake_case JSON into one camelCase trust block:
```
agent UUID ──────────▶ GET /api/agents/{id}/reputation ─▶ WalletReputation
Solana asset (base58) ─▶ GET /api/agents/solana-reputation ─▶ SolanaReputation
- (—) ─────────────────▶ GET /api/reputation/leaderboard ─▶ Leaderboard
+ (none) ──────────────▶ GET /api/reputation/leaderboard ─▶ Leaderboard
chainId + agentId ────▶ GET /api/erc8004/validation ─▶ ValidationRead
attest, Solana asset ─▶ POST /api/agents/solana-validate ┐
attest, EVM agentId ──▶ POST /api/erc8004/validate ┴─▶ AttestReceipt (signed on-chain)
```
-- **Wallet-trust** (UUID) is the platform's non-gameable score — the same number
+- **Wallet-trust** (UUID) is the platform's non-gameable score, the same number
the badge and the leaderboard show, computed from real ledger + chain activity.
- **On-chain aggregate** (asset) rolls up the agent's attestations, stake,
validations, disputes, and activity straight from the indexer.
@@ -284,7 +284,7 @@ normalises the snake_case JSON into one camelCase trust block:
attestation is a signed on-chain record of the agent's GLB passing the platform
validator, keyed by asset/agentId so a retry re-records idempotently.
-Everything is a live read/write — a brand-new agent returns `score: null` /
+Everything is a live read/write; a brand-new agent returns `score: null` /
zeroed feedback totals, never an invented number.
## Chains
@@ -293,12 +293,12 @@ zeroed feedback totals, never an invented number.
(`validation()` reads and the EVM `attest()` write) understands. Select one by
name (`'base'`, `'arbitrum'`, …) or numeric id; the default is **Base**.
-**Mainnets** — Base (8453), Arbitrum One (42161), BNB Chain (56), Ethereum (1),
+**Mainnets**: Base (8453), Arbitrum One (42161), BNB Chain (56), Ethereum (1),
Optimism (10), Polygon (137), Avalanche (43114), Gnosis (100), Fantom (250),
Celo (42220), Linea (59144), Scroll (534352), Mantle (5000), zkSync Era (324),
Moonbeam (1284).
-**Testnets** — BSC Testnet (97), Base Sepolia (84532), Arbitrum Sepolia (421614),
+**Testnets**: BSC Testnet (97), Base Sepolia (84532), Arbitrum Sepolia (421614),
Ethereum Sepolia (11155111), Optimism Sepolia (11155420), Polygon Amoy (80002),
Avalanche Fuji (43113).
@@ -314,7 +314,7 @@ typed `ThreeWsError` with a stable `code` and the HTTP `status`:
| `invalid_input` | `agent` isn't a UUID or Solana asset (or `agentId` isn't a uint). | Pass a valid identifier. |
| `unsupported_chain` | Unknown chain name / id in `validation()` / `attest()`. | Use a [supported chain](#chains). |
| `score: null` / `isNew: true` | The agent has no track record yet. | Surface "new agent", not a zero score. |
-| `feedback.total === 0` | No on-chain attestations for the asset yet. | Render an empty state — that's the truth. |
+| `feedback.total === 0` | No on-chain attestations for the asset yet. | Render an empty state; that's the truth. |
| `unauthorized` (401) | `attest()` without a session or token. | Sign in, or pass an `avatars:write` `apiKey`. |
| `payment_required` (402) | A gated lane needs payment. | Handle the thrown `PaymentRequiredError` (`.accepts`). |
| `rate_limited` (429) | Too many requests. | Honour `retryAfter` on the error. |
@@ -324,7 +324,7 @@ An empty registry or a missing attestation returns a real empty result
## Examples
-**Marketplace gate** — only list agents above a trust floor:
+**Marketplace gate**: only list agents above a trust floor:
```js
import { reputation } from '@three-ws/reputation';
@@ -355,15 +355,15 @@ import { leaderboard } from '@three-ws/reputation';
const { agents } = await leaderboard({ limit: 10 });
agents.forEach((a) =>
- console.log(`#${a.rank} ${a.name} — ${a.tierLabel} (${a.score})`),
+ console.log(`#${a.rank} ${a.name}: ${a.tierLabel} (${a.score})`),
);
```
## Related
-- [`@three-ws/avatar-schema`](https://www.npmjs.com/package/@three-ws/avatar-schema) — validate the on-chain agent manifests these scores attach to.
-- [`@three-ws/forge`](https://www.npmjs.com/package/@three-ws/forge) — generate the rig-ready GLB an agent identity wears.
-- [`@three-ws/x402-fetch`](https://www.npmjs.com/package/@three-ws/x402-fetch) — auto-pay the `agent_reputation` MCP tool's $0.01 lane.
+- [`@three-ws/avatar-schema`](https://www.npmjs.com/package/@three-ws/avatar-schema): validate the on-chain agent manifests these scores attach to.
+- [`@three-ws/forge`](https://www.npmjs.com/package/@three-ws/forge): generate the rig-ready GLB an agent identity wears.
+- [`@three-ws/x402-fetch`](https://www.npmjs.com/package/@three-ws/x402-fetch): auto-pay the `agent_reputation` MCP tool's $0.01 lane.
---
diff --git a/packages/reputation/package.json b/packages/reputation/package.json
index 0abf0ed6e..435b77908 100644
--- a/packages/reputation/package.json
+++ b/packages/reputation/package.json
@@ -1,7 +1,7 @@
{
"name": "@three-ws/reputation",
"version": "0.1.1",
- "description": "Read ERC-8004 agent trust scores and attest agent-to-agent feedback on-chain, in one import. The three.ws reputation SDK — canonical Identity + Reputation registries on Base and 14 other chains.",
+ "description": "Read ERC-8004 agent trust scores and attest agent-to-agent feedback on-chain, in one import. The three.ws reputation SDK: canonical Identity + Reputation registries on Base and 14 other chains.",
"type": "module",
"main": "./src/index.js",
"module": "./src/index.js",
diff --git a/packages/reputation/src/http.js b/packages/reputation/src/http.js
index 6398a3e53..339c074eb 100644
--- a/packages/reputation/src/http.js
+++ b/packages/reputation/src/http.js
@@ -20,7 +20,7 @@ export class ThreeWsError extends Error {
/**
* Thrown on HTTP 402. The endpoint wants payment before it will do the work.
- * `accepts` is the x402 challenge (asset/amount/network/payTo) when present —
+ * `accepts` is the x402 challenge (asset/amount/network/payTo) when present;
* pass a payment-aware `fetch` (e.g. @three-ws/x402-fetch) to settle it
* automatically, or read `accepts` and pay it yourself.
*/
@@ -63,7 +63,7 @@ export function createHttp(opts = {}) {
const baseUrl = resolveBaseUrl(opts.baseUrl);
const fetchImpl = opts.fetch || (typeof globalThis !== 'undefined' ? globalThis.fetch : undefined);
if (typeof fetchImpl !== 'function') {
- throw new ThreeWsError('No fetch implementation available — run on Node 18+ or pass { fetch }.', { code: 'no_fetch' });
+ throw new ThreeWsError('No fetch implementation available: run on Node 18+ or pass { fetch }.', { code: 'no_fetch' });
}
const baseHeaders = { accept: 'application/json', ...(opts.headers || {}) };
if (opts.apiKey) baseHeaders.authorization = `Bearer ${opts.apiKey}`;
diff --git a/packages/reputation/src/index.d.ts b/packages/reputation/src/index.d.ts
index 1e0c814a1..e31acd573 100644
--- a/packages/reputation/src/index.d.ts
+++ b/packages/reputation/src/index.d.ts
@@ -94,7 +94,7 @@ export interface SolanaReputation {
export type ReputationResult = WalletReputation | SolanaReputation;
export interface LeaderboardOptions {
- /** 1–50, default 20. */
+ /** 1-50, default 20. */
limit?: number;
signal?: AbortSignal;
}
diff --git a/packages/reputation/src/index.js b/packages/reputation/src/index.js
index f7255f1df..68e5b4659 100644
--- a/packages/reputation/src/index.js
+++ b/packages/reputation/src/index.js
@@ -1,4 +1,4 @@
-// @three-ws/reputation — read ERC-8004 agent trust scores and attest
+// @three-ws/reputation: read ERC-8004 agent trust scores and attest
// agent-to-agent feedback on-chain, in one import. Zero runtime deps: every
// on-chain read/write is done server-side by the three.ws platform endpoints
// this client wraps (the same registries the `agent_reputation` MCP tool reads),
@@ -129,7 +129,7 @@ export function createReputation(options = {}) {
/**
* Read the latest on-chain ERC-8004 validation attestation for an agent
- * (GET /api/erc8004/validation) — the walletless identity-bearing read that
+ * (GET /api/erc8004/validation): the walletless identity-bearing read that
* powers the "Validated" badge. `chain` selects the network.
*/
async function validation(chainId, agentId, opts = {}) {
diff --git a/packages/reputation/test/reputation.test.js b/packages/reputation/test/reputation.test.js
index d485ca013..70da93231 100644
--- a/packages/reputation/test/reputation.test.js
+++ b/packages/reputation/test/reputation.test.js
@@ -3,7 +3,7 @@ import assert from 'node:assert/strict';
import { createReputation, SUPPORTED_CHAINS, ThreeWsError } from '../src/index.js';
// A scripted fetch double: each call shifts the next queued response and records
-// the request. No network, no real endpoints — we assert on request shaping and
+// the request. No network, no real endpoints, we assert on request shaping and
// response parsing, which is all the SDK is responsible for.
function stubFetch(responses) {
const calls = [];
@@ -23,7 +23,7 @@ function stubFetch(responses) {
return { fetch, calls };
}
-// A synthetic Solana asset address (base58) — never a real third-party mint.
+// A synthetic Solana asset address (base58): never a real third-party mint.
const SYNTH_ASSET = 'THREEsynthetic1111111111111111111111111111';
const AGENT_UUID = '7b9a4f30-2d11-4e2d-9d12-1cdb1f6a3a55';
diff --git a/packages/retarget/README.md b/packages/retarget/README.md
index 32e0cc1bf..72c5d00e1 100644
--- a/packages/retarget/README.md
+++ b/packages/retarget/README.md
@@ -1,10 +1,10 @@
# @three-ws/retarget
-> **Retarget animations onto any humanoid GLB.** One package that maps any rig's bone names onto a canonical skeleton, retargets clips onto it with rest-pose and hip-axis correction, and drives them with a crossfading runtime — so no humanoid ever freezes in a T-pose.
+> **Retarget animations onto any humanoid GLB.** One package that maps any rig's bone names onto a canonical skeleton, retargets clips onto it with rest-pose and hip-axis correction, and drives them with a crossfading runtime, so no humanoid ever freezes in a T-pose.
-This is the animation engine behind every avatar on [three.ws](https://three.ws) — the same code that lets [`@three-ws/walk`](https://www.npmjs.com/package/@three-ws/walk), [`@three-ws/tour`](https://www.npmjs.com/package/@three-ws/tour), and the platform's viewers animate a GLB from **any** source without a curated rig allowlist. Extracted here so you can use it with plain Three.js.
+This is the animation engine behind every avatar on [three.ws](https://three.ws): the same code that lets [`@three-ws/walk`](https://www.npmjs.com/package/@three-ws/walk), [`@three-ws/tour`](https://www.npmjs.com/package/@three-ws/tour), and the platform's viewers animate a GLB from **any** source without a curated rig allowlist. Extracted here so you can use it with plain Three.js.
-**Rig conventions it understands out of the box:** Mixamo (`mixamorig:LeftArm`), Avaturn / Wolf3D, VRM 0.x / VRoid (`J_Bip_L_UpperArm`), VRM 1.0, Daz/Genesis, MakeHuman, Unreal mannequin, HumanIK/Maya, Blender `.L`/`.R`, bare `shoulderL`-style names, and casing/separator variants of all of them. A rig that genuinely isn't a skinned humanoid is detected and refused cleanly — never half-animated.
+**Rig conventions it understands out of the box:** Mixamo (`mixamorig:LeftArm`), Avaturn / Wolf3D, VRM 0.x / VRoid (`J_Bip_L_UpperArm`), VRM 1.0, Daz/Genesis, MakeHuman, Unreal mannequin, HumanIK/Maya, Blender `.L`/`.R`, bare `shoulderL`-style names, and casing/separator variants of all of them. A rig that genuinely isn't a skinned humanoid is detected and refused cleanly, never half-animated.
---
@@ -16,7 +16,7 @@ npm install @three-ws/retarget three
`three` is a peer dependency (>= 0.150).
-## Quick start — play a clip on a rig it was never made for
+## Quick start: play a clip on a rig it was never made for
```js
import { AnimationMixer } from 'three';
@@ -72,7 +72,7 @@ if (manager.supportsCanonicalClips()) {
manager.update(deltaSeconds);
```
-The clip URLs above are real — three.ws serves its shared clip library with open CORS, so the snippet runs as-is. Ship your own clips in production if you need guaranteed availability.
+The clip URLs above are real: three.ws serves its shared clip library with open CORS, so the snippet runs as-is. Ship your own clips in production if you need guaranteed availability.
## API surface
@@ -83,7 +83,7 @@ The clip URLs above are real — three.ws serves its shared clip library with op
| `canonicalizeBoneName(name)` | One bone name → canonical name (or passthrough) |
| `canonicalizeJointNodes(json)` | Rename joints across a parsed glTF JSON tree |
| `canonicalizeArmatureOrientation(json)` | Normalize armature up-axis quirks |
-| `canonicalizeGLBBones(arrayBuffer)` | The whole pipeline on a GLB binary — byte-exact when nothing needs changing |
+| `canonicalizeGLBBones(arrayBuffer)` | The whole pipeline on a GLB binary, byte-exact when nothing needs changing |
**Retarget** (`animation-retarget`):
| Export | What it does |
@@ -101,19 +101,19 @@ The clip URLs above are real — three.ws serves its shared clip library with op
| `AnimationManager` | attach/detach, `setAnimationDefs`, `loadAll`, `crossfadeTo`, overlays, `supportsCanonicalClips()`, fallen-pose guards, `update(dt)` |
| `measureHipsTiltDeg(clip, model, map)` | Diagnostic: how far a retargeted clip tilts the hips |
-**Reference data:** `CANONICAL_REST`, `CANONICAL_REST_WORLD` — the canonical rig's rest quaternions the corrections are computed against.
+**Reference data:** `CANONICAL_REST` and `CANONICAL_REST_WORLD` hold the canonical rig's rest quaternions the corrections are computed against.
## How it works
Retargeting quality comes from doing the correction in world space: for each bone the engine computes `q' = L · q · R`, where `L`/`R` are derived from the difference between the clip's canonical rest pose and the target rig's actual rest pose (so an A-pose character plays T-pose clips without shoulder skew), plus a hip-space fix for rigs whose hips aren't Y-up. Coverage below 50% of the canonical bones aborts the retarget instead of producing a half-puppet.
-Hit a skeleton convention it doesn't know? Add the bone-name mapping to the canonicalizer and it works everywhere at once — that's the design: **map names once, never maintain a rig allowlist.**
+Hit a skeleton convention it doesn't know? Add the bone-name mapping to the canonicalizer and it works everywhere at once; that's the design: **map names once, never maintain a rig allowlist.**
## Related packages
-- [`@three-ws/walk`](https://www.npmjs.com/package/@three-ws/walk) — a walking avatar companion for any site, built on this engine
-- [`@three-ws/tour`](https://www.npmjs.com/package/@three-ws/tour) — guided site tours with a walking narrator
-- [`@three-ws/page-agent`](https://www.npmjs.com/package/@three-ws/page-agent) — a docked, talking page narrator
+- [`@three-ws/walk`](https://www.npmjs.com/package/@three-ws/walk): a walking avatar companion for any site, built on this engine
+- [`@three-ws/tour`](https://www.npmjs.com/package/@three-ws/tour): guided site tours with a walking narrator
+- [`@three-ws/page-agent`](https://www.npmjs.com/package/@three-ws/page-agent): a docked, talking page narrator
## License
diff --git a/packages/retarget/build.mjs b/packages/retarget/build.mjs
index bb6cd4d35..83784b4e1 100644
--- a/packages/retarget/build.mjs
+++ b/packages/retarget/build.mjs
@@ -3,7 +3,7 @@
// =======================================================================
// The entry re-exports the engine from the monorepo source; esbuild bundles
// those files (and the tiny shared logger) into dist/index.mjs, leaving only
-// `three` external as the peer dependency — the same publish pattern as
+// `three` external as the peer dependency, the same publish pattern as
// @three-ws/walk.
import { build } from 'esbuild';
diff --git a/packages/retarget/package.json b/packages/retarget/package.json
index 276dc0deb..43af11dba 100644
--- a/packages/retarget/package.json
+++ b/packages/retarget/package.json
@@ -1,7 +1,7 @@
{
"name": "@three-ws/retarget",
"version": "0.1.1",
- "description": "Retarget animations onto any humanoid GLB. Canonicalizes bone names from Mixamo, Avaturn, VRM/VRoid, Daz/Genesis, MakeHuman, Unreal, and Blender rigs onto one canonical skeleton, retargets clips with rest-pose and hip-axis correction, and drives them with a crossfading runtime — so no rig ever freezes in a T-pose.",
+ "description": "Retarget animations onto any humanoid GLB. Canonicalizes bone names from Mixamo, Avaturn, VRM/VRoid, Daz/Genesis, MakeHuman, Unreal, and Blender rigs onto one canonical skeleton, retargets clips with rest-pose and hip-axis correction, and drives them with a crossfading runtime, so no rig ever freezes in a T-pose.",
"keywords": [
"three.js",
"animation",
diff --git a/packages/retarget/src/index.js b/packages/retarget/src/index.js
index 8727080e7..049889324 100644
--- a/packages/retarget/src/index.js
+++ b/packages/retarget/src/index.js
@@ -1,18 +1,18 @@
-// @three-ws/retarget — public entry.
+// @three-ws/retarget: public entry.
// ===================================
// Re-exports the platform's rig-agnostic humanoid animation engine from the
// monorepo source (the seam walk-sdk's PROMOTION NOTE designates); the publish
// build bundles everything into dist/ so npm consumers get one standalone
// package whose only dependency is their own copy of Three.js.
//
-// 1. canonicalize — map any humanoid rig's bone names (Mixamo, Avaturn,
+// 1. canonicalize: map any humanoid rig's bone names (Mixamo, Avaturn,
// VRM/VRoid, VRM 1.0, Daz/Genesis, MakeHuman, Unreal, Blender `.L`,
// simple `shoulderL` rigs, …) onto the frozen canonical bone set, either
// on a parsed glTF JSON tree or byte-exactly inside a GLB ArrayBuffer.
-// 2. retarget — play any canonical-space AnimationClip on that rig, with
+// 2. retarget: play any canonical-space AnimationClip on that rig, with
// rest-pose (A/T-pose) correction, hip up-axis fixes, and coverage gates
// so a non-humanoid never half-animates.
-// 3. drive — the AnimationManager runtime that loads clip libraries,
+// 3. drive: the AnimationManager runtime that loads clip libraries,
// crossfades, layers overlays, and refuses fallen/broken poses.
export {
diff --git a/packages/retarget/test/exports.test.mjs b/packages/retarget/test/exports.test.mjs
index 7e061557c..30f9409e8 100644
--- a/packages/retarget/test/exports.test.mjs
+++ b/packages/retarget/test/exports.test.mjs
@@ -1,11 +1,11 @@
-// @three-ws/retarget — published-surface smoke tests.
+// @three-ws/retarget: published-surface smoke tests.
// =====================================================
// The engine's deep behavioral coverage lives in the monorepo suite
// (tests/glb-canonicalize.test.js, tests/animation-retarget.test.js,
-// tests/animation-upright-invariant.test.js — ~2.8k lines, run by root
+// tests/animation-upright-invariant.test.js, ~2.8k lines, run by root
// vitest). This file guards the PACKAGE: the bundled dist exposes the full
// export surface and the canonicalizer actually resolves real-world rig
-// names. `node --test`, runs against dist/ (build first — prepublishOnly does).
+// names. `node --test`, runs against dist/ (build first; prepublishOnly does).
import { test } from 'node:test';
import assert from 'node:assert/strict';
diff --git a/packages/scene-mcp/README.md b/packages/scene-mcp/README.md
index 97cb84c34..ecbb859ea 100644
--- a/packages/scene-mcp/README.md
+++ b/packages/scene-mcp/README.md
@@ -4,7 +4,7 @@
<h1 align="center">@three-ws/scene-mcp</h1>
-<p align="center"><strong>Speak 3D worlds into being — turn one sentence into a placed diorama from any AI agent.</strong></p>
+<p align="center"><strong>Speak 3D worlds into being: turn one sentence into a placed diorama from any AI agent.</strong></p>
<p align="center">
<a href="https://www.npmjs.com/package/@three-ws/scene-mcp"><img alt="npm" src="https://img.shields.io/npm/v/@three-ws/scene-mcp?logo=npm&color=cb3837"></a>
@@ -16,9 +16,9 @@
---
-> A [Model Context Protocol](https://modelcontextprotocol.io) server that gives any AI assistant the three.ws **diorama** pipeline over stdio. Describe a world in a sentence and `compose_scene` returns a placed plan — mood, palette, ground, and a set of single-object forge prompts — ready to build into an orbitable 3D scene. `build_world` collapses the whole pipeline (compose → forge every object → merge) into one call for agents with no browser to drive it. `export_scene` merges an already-forged diorama into one glTF 2.0 binary — every object a named, selectable node, plus ground and mood-tuned lighting — ready to open in Scene Studio. Browse saved worlds with `get_scene` and `list_scenes`.
+> A [Model Context Protocol](https://modelcontextprotocol.io) server that gives any AI assistant the three.ws **diorama** pipeline over stdio. Describe a world in a sentence and `compose_scene` returns a placed plan (mood, palette, ground, and a set of single-object forge prompts), ready to build into an orbitable 3D scene. `build_world` collapses the whole pipeline (compose → forge every object → merge) into one call for agents with no browser to drive it. `export_scene` merges an already-forged diorama into one glTF 2.0 binary (every object a named, selectable node, plus ground and mood-tuned lighting), ready to open in Scene Studio. Browse saved worlds with `get_scene` and `list_scenes`.
-This is the companion to the three.ws avatar one-liner: avatars give you a *character*; scenes give you a *world*. No API key, no signer, no payment — every call hits the public `/api/diorama` endpoint.
+This is the companion to the three.ws avatar one-liner: avatars give you a *character*; scenes give you a *world*. No API key, no signer, no payment: every call hits the public `/api/diorama` endpoint.
## Install
@@ -63,25 +63,25 @@ npx -y @modelcontextprotocol/inspector npx @three-ws/scene-mcp
| Tool | Type | What it does |
| --------------- | ----------- | ----------------------------------------------------------------------------------------------------------------- |
-| `compose_scene` | generative | One sentence → a diorama plan: title, mood, palette, ground, and 2–8 placed single-object forge prompts. |
+| `compose_scene` | generative | One sentence → a diorama plan: title, mood, palette, ground, and 2-8 placed single-object forge prompts. |
| `build_world` | generative | One sentence → a fully forged, exported 3D world in one call: compose → forge every object (free lane) → merge into one GLB. No browser needed; can take a couple of minutes. |
-| `export_scene` | generative | Merge an already-forged diorama (objects with real `glbUrl`s) into one glTF 2.0 binary — named, selectable nodes plus ground and mood-tuned lighting — ready for Scene Studio. |
+| `export_scene` | generative | Merge an already-forged diorama (objects with real `glbUrl`s) into one glTF 2.0 binary (named, selectable nodes plus ground and mood-tuned lighting), ready for Scene Studio. |
| `get_scene` | read-only | Fetch a saved world by id (with GLB URLs) and its orbitable viewer URL. |
| `list_scenes` | read-only | Browse the recent or featured diorama gallery. |
-`compose_scene` returns only the *plan* — no meshes are forged and nothing is saved. Forge each object via `/api/forge`, then either call `export_scene` yourself or `POST /api/diorama {action:"save"}` to mint a shareable permalink at `https://three.ws/diorama?id=…`. `build_world` does the compose-forge-export sequence for you in a single tool call. Both `build_world` and `export_scene` degrade gracefully on partial failure: objects that never forge are skipped and named in the response's `skipped` array — the rest of the world still exports.
+`compose_scene` returns only the *plan*: no meshes are forged and nothing is saved. Forge each object via `/api/forge`, then either call `export_scene` yourself or `POST /api/diorama {action:"save"}` to mint a shareable permalink at `https://three.ws/diorama?id=…`. `build_world` does the compose-forge-export sequence for you in a single tool call. Both `build_world` and `export_scene` degrade gracefully on partial failure: objects that never forge are skipped and named in the response's `skipped` array; the rest of the world still exports.
### Input parameters
-**`compose_scene`** — `prompt` (required, 3–1024 chars: one sentence describing the world).
+**`compose_scene`**: `prompt` (required, 3-1024 chars: one sentence describing the world).
-**`build_world`** — `prompt` (required, 3–1024 chars: one sentence describing the world).
+**`build_world`**: `prompt` (required, 3-1024 chars: one sentence describing the world).
-**`export_scene`** — `diorama` (required: the diorama object returned by `compose_scene`/`get_scene`, with each object to include carrying `status:"ready"` and a real `glbUrl`).
+**`export_scene`**: `diorama` (required: the diorama object returned by `compose_scene`/`get_scene`, with each object to include carrying `status:"ready"` and a real `glbUrl`).
-**`get_scene`** — `id` (required: the diorama id returned at save time).
+**`get_scene`**: `id` (required: the diorama id returned at save time).
-**`list_scenes`** — `list` (`recent` | `featured`, default `recent`), `limit` (1–50, default 24).
+**`list_scenes`**: `list` (`recent` | `featured`, default `recent`), `limit` (1-50, default 24).
## Example
@@ -106,7 +106,7 @@ npx -y @modelcontextprotocol/inspector npx @three-ws/scene-mcp
```
```jsonc
-// build_world — the whole pipeline in one call
+// build_world: the whole pipeline in one call
> { "prompt": "a neon alley with a food cart and two streetlights" }
{
"ok": true,
@@ -137,13 +137,13 @@ npx -y @modelcontextprotocol/inspector npx @three-ws/scene-mcp
- Homepage: https://three.ws
- Changelog: https://three.ws/changelog
- Issues: https://github.com/nirholas/three.ws/issues
-- License: Apache-2.0 — see [LICENSE](./LICENSE)
+- License: all rights reserved, see [LICENSE](./LICENSE)
---
<p align="center">
<sub>
- Part of the <a href="https://three.ws">three.ws</a> SDK suite — 3D AI agents, on-chain identity, and agent payments.<br/>
+ Part of the <a href="https://three.ws">three.ws</a> SDK suite: 3D AI agents, on-chain identity, and agent payments.<br/>
<a href="https://three.ws">Website</a> · <a href="https://three.ws/changelog">Changelog</a> · <a href="https://github.com/nirholas/three.ws">GitHub</a>
</sub>
</p>
diff --git a/packages/scene-mcp/src/config.js b/packages/scene-mcp/src/config.js
index c442b22c3..c7e01467e 100644
--- a/packages/scene-mcp/src/config.js
+++ b/packages/scene-mcp/src/config.js
@@ -1,7 +1,7 @@
// Centralized env + HTTP base for the scene MCP.
//
// This server is a thin, read-mostly wrapper over the PUBLIC three.ws HTTP
-// API (/api/diorama). It signs nothing and holds no secret — the only knobs
+// API (/api/diorama). It signs nothing and holds no secret; the only knobs
// are which deployment to talk to and how long to wait. Nothing about how a
// world is composed is baked in here; it all comes from the live endpoint.
diff --git a/packages/scene-mcp/src/index.js b/packages/scene-mcp/src/index.js
index 08874ab58..0c6e2b04e 100755
--- a/packages/scene-mcp/src/index.js
+++ b/packages/scene-mcp/src/index.js
@@ -1,20 +1,20 @@
#!/usr/bin/env node
-// @three-ws/scene-mcp — MCP server entry point.
+// @three-ws/scene-mcp: MCP server entry point.
//
// Gives any AI assistant the three.ws diorama pipeline over stdio:
-// • compose_scene — one sentence → a placed 3D diorama plan (LLM-composed)
-// • get_scene — fetch a saved world by id
-// • list_scenes — browse the recent / featured gallery
-// • export_scene — merge an already-forged diorama into one GLB scene
-// • build_world — one sentence → a fully forged, exported world (no browser)
+// • compose_scene: one sentence → a placed 3D diorama plan (LLM-composed)
+// • get_scene: fetch a saved world by id
+// • list_scenes: browse the recent / featured gallery
+// • export_scene: merge an already-forged diorama into one GLB scene
+// • build_world: one sentence → a fully forged, exported world (no browser)
//
// A thin wrapper over the PUBLIC three.ws API (/api/diorama). No keys, no
-// signer, no payment — point THREE_WS_BASE at a deployment and go.
+// signer, no payment: point THREE_WS_BASE at a deployment and go.
//
// Run standalone:
// node packages/scene-mcp/src/index.js
//
-// Or wire into Claude Code / Cursor — see README.md.
+// Or wire into Claude Code / Cursor (see README.md).
import { realpathSync } from 'node:fs';
import { createRequire } from 'node:module';
@@ -29,7 +29,7 @@ import { def as listScenes } from './tools/list-scenes.js';
import { def as exportScene } from './tools/export-scene.js';
import { def as buildWorld } from './tools/build-world.js';
-// Single source of truth for the advertised server version — package.json.
+// Single source of truth for the advertised server version: package.json.
const require = createRequire(import.meta.url);
const { version: PKG_VERSION } = require('../package.json');
@@ -46,16 +46,16 @@ export function buildServer() {
{
capabilities: { tools: {} },
instructions:
- 'three.ws Scene MCP — speak 3D worlds into being. compose_scene turns one short sentence ' +
- 'into a diorama PLAN (title, mood, palette, ground, and 2–8 placed single-object forge prompts) ' +
+ 'three.ws Scene MCP: speak 3D worlds into being. compose_scene turns one short sentence ' +
+ 'into a diorama PLAN (title, mood, palette, ground, and 2-8 placed single-object forge prompts) ' +
'using the platform free-first LLM chain; nothing is saved and no meshes are forged yet. ' +
'get_scene fetches a previously saved world by id (with GLB URLs and an orbitable viewer link). ' +
'list_scenes browses the recent or featured gallery. export_scene merges an already-forged ' +
- 'diorama into ONE glTF 2.0 binary — every object a named, selectable node, plus a real ground ' +
- 'disc and mood-tuned lighting — ready to open in Scene Studio. build_world runs the whole ' +
+ 'diorama into ONE glTF 2.0 binary (every object a named, selectable node, plus a real ground ' +
+ 'disc and mood-tuned lighting), ready to open in Scene Studio. build_world runs the whole ' +
'pipeline (compose → forge every object → export) server-side from one sentence, for callers ' +
'with no browser to drive the progressive flow; it can take a couple of minutes. All data comes ' +
- 'live from the public three.ws /api/diorama endpoint — no API key, signer, or payment required.',
+ 'live from the public three.ws /api/diorama endpoint; no API key, signer, or payment required.',
},
);
diff --git a/packages/scene-mcp/src/lib/api.js b/packages/scene-mcp/src/lib/api.js
index 68ead13b1..7109f2d1b 100644
--- a/packages/scene-mcp/src/lib/api.js
+++ b/packages/scene-mcp/src/lib/api.js
@@ -1,4 +1,4 @@
-// Real HTTP access to the three.ws API. No mocks, no fixtures — every call is
+// Real HTTP access to the three.ws API. No mocks, no fixtures: every call is
// a live request to THREE_WS_BASE. Errors are normalized into a single shape so
// tool handlers can surface a clean message + status to the MCP client.
@@ -10,7 +10,7 @@ import { THREE_WS_BASE, HTTP_TIMEOUT_MS, USER_AGENT } from '../config.js';
* @param {string} path Endpoint path beginning with `/` (e.g. `/api/diorama`).
* @param {{ method?: string, query?: Record<string, unknown>, body?: unknown, timeoutMs?: number }} [opts]
* `timeoutMs` overrides the default per-call timeout (THREE_WS_TIMEOUT_MS) for
- * calls known to run long server-side work — e.g. a full compose+forge+export
+ * calls known to run long server-side work, e.g. a full compose+forge+export
* build, which chains several real generations. Every other call is unaffected.
* @returns {Promise<any>} Parsed JSON response.
* @throws {Error} with `.code` ('timeout' | 'network_error' | 'upstream_error'),
diff --git a/packages/scene-mcp/src/tools/build-world.js b/packages/scene-mcp/src/tools/build-world.js
index e22ef2f9a..f32b87a2e 100644
--- a/packages/scene-mcp/src/tools/build-world.js
+++ b/packages/scene-mcp/src/tools/build-world.js
@@ -1,4 +1,4 @@
-// `build_world` — one sentence → a fully forged, exported 3D world. No browser.
+// `build_world`: one sentence → a fully forged, exported 3D world. No browser.
//
// Wraps POST /api/diorama {action:'build'}. Runs the ENTIRE diorama pipeline
// server-side in one call: compose the plan (free-first LLM chain), forge
@@ -8,7 +8,7 @@
// agents/MCP clients that have no browser to drive that flow.
//
// Real generation work (an LLM completion plus N free-lane 3D forges), so this
-// can legitimately take minutes for a full object set — the tool requests a
+// can legitimately take minutes for a full object set, so the tool requests a
// long timeout accordingly. A forge failure on some objects never fails the
// whole call: the returned diorama reflects exactly what forged, and `skipped`
// names the rest, so you always get the best real world the platform could
@@ -26,7 +26,7 @@ export const def = {
title: 'Build a complete 3D world from one sentence (no browser needed)',
annotations: { readOnlyHint: false, destructiveHint: false, idempotentHint: false, openWorldHint: true },
description:
- 'One sentence → a fully forged, exported 3D world, entirely server-side: composes the scene plan, forges every object on the free text→3D lane, and merges the result into one GLB with named, selectable nodes plus ground and lighting. Returns the populated diorama, the merged GLB URL, and a ready-to-open Scene Studio link. This is the whole progressive /diorama browser flow collapsed into one call — it can take a couple of minutes for a full object set. Partial forges are not failures: whatever forged is exported, and `skipped` names anything that did not.',
+ 'One sentence → a fully forged, exported 3D world, entirely server-side: composes the scene plan, forges every object on the free text→3D lane, and merges the result into one GLB with named, selectable nodes plus ground and lighting. Returns the populated diorama, the merged GLB URL, and a ready-to-open Scene Studio link. This is the whole progressive /diorama browser flow collapsed into one call; it can take a couple of minutes for a full object set. Partial forges are not failures: whatever forged is exported, and `skipped` names anything that did not.',
inputSchema: {
prompt: z
.string()
diff --git a/packages/scene-mcp/src/tools/compose-scene.js b/packages/scene-mcp/src/tools/compose-scene.js
index d78525e31..9f4baea9b 100644
--- a/packages/scene-mcp/src/tools/compose-scene.js
+++ b/packages/scene-mcp/src/tools/compose-scene.js
@@ -1,8 +1,8 @@
-// `compose_scene` — turn one sentence into a placed 3D diorama plan.
+// `compose_scene`: turn one sentence into a placed 3D diorama plan.
//
// Wraps POST /api/diorama {action:'compose'}. The platform's free-first LLM
// chain decomposes the sentence into a mood/palette/ground plus a set of
-// single-object forge prompts with positions — the same plan the three.ws
+// single-object forge prompts with positions, the same plan the three.ws
// diorama page then forges into meshes. Returns the plan; nothing is persisted.
import { z } from 'zod';
@@ -15,11 +15,11 @@ export const def = {
name: 'compose_scene',
title: 'Compose a 3D diorama plan from one sentence',
// Not a pure read: it drives a server-side LLM completion. It mutates no
- // state, so it is not destructive — but annotation-aware clients should
+ // state, so it is not destructive, but annotation-aware clients should
// still surface it rather than silently auto-running inference.
annotations: { readOnlyHint: false, destructiveHint: false, idempotentHint: false, openWorldHint: true },
description:
- 'Compose a tiny 3D diorama from one short sentence. Returns a PLAN: an evocative title, a mood (dawn/day/dusk/night), ground + island type, a color palette, and 2–8 placed objects — each with a single-object forge prompt, position, scale and rotation. No meshes are generated yet and nothing is saved; feed the plan to the three.ws diorama page (or /api/forge per object) to build the world. Read-mostly but runs live LLM inference.',
+ 'Compose a tiny 3D diorama from one short sentence. Returns a PLAN: an evocative title, a mood (dawn/day/dusk/night), ground + island type, a color palette, and 2-8 placed objects, each with a single-object forge prompt, position, scale and rotation. No meshes are generated yet and nothing is saved; feed the plan to the three.ws diorama page (or /api/forge per object) to build the world. Read-mostly but runs live LLM inference.',
inputSchema: {
prompt: z
.string()
diff --git a/packages/scene-mcp/src/tools/export-scene.js b/packages/scene-mcp/src/tools/export-scene.js
index 947ef1828..50de5bf1c 100644
--- a/packages/scene-mcp/src/tools/export-scene.js
+++ b/packages/scene-mcp/src/tools/export-scene.js
@@ -1,12 +1,12 @@
-// `export_scene` — merge an already-forged diorama into one exportable GLB.
+// `export_scene`: merge an already-forged diorama into one exportable GLB.
//
// Wraps POST /api/diorama {action:'export'}. Takes a diorama that already has
-// forged objects (glbUrl per object — from compose_scene + your own forging,
+// forged objects (glbUrl per object, from compose_scene + your own forging,
// or from a world fetched with get_scene) and merges it server-side into ONE
// glTF 2.0 binary: every object as a named node, plus a real ground disc and
// KHR_lights_punctual lighting tuned to the diorama's mood. The result opens
// cleanly in three.ws Scene Studio (or any glTF viewer) with every object
-// individually selectable. Partial worlds export fine — unforged/failed
+// individually selectable. Partial worlds export fine; unforged/failed
// objects are just skipped and reported back, never a hard failure.
import { z } from 'zod';
@@ -20,18 +20,18 @@ export const def = {
title: 'Export a forged diorama as one GLB scene',
annotations: { readOnlyHint: false, destructiveHint: false, idempotentHint: false, openWorldHint: true },
description:
- 'Merge an already-forged diorama (objects with real glbUrl values, e.g. from your own forging after compose_scene, or fetched via get_scene) into ONE glTF 2.0 binary: every object becomes a named, selectable node, plus a real ground disc and mood-tuned lighting. Returns the merged GLB URL and a ready-to-open three.ws Scene Studio link. Objects that never forged are skipped and reported, not treated as a failure — a partial world still exports. Requires the target deployment to have object storage configured (three.ws does).',
+ 'Merge an already-forged diorama (objects with real glbUrl values, e.g. from your own forging after compose_scene, or fetched via get_scene) into ONE glTF 2.0 binary: every object becomes a named, selectable node, plus a real ground disc and mood-tuned lighting. Returns the merged GLB URL and a ready-to-open three.ws Scene Studio link. Objects that never forged are skipped and reported rather than treated as a failure; a partial world still exports. Requires the target deployment to have object storage configured (three.ws does).',
inputSchema: {
diorama: z
.record(z.string(), z.any())
.describe(
- 'The diorama object to export — same shape returned by compose_scene/get_scene, with each object you want included carrying status:"ready" and a real glbUrl.',
+ 'The diorama object to export, same shape returned by compose_scene/get_scene, with each object you want included carrying status:"ready" and a real glbUrl.',
),
},
async handler(args) {
const diorama = args?.diorama;
if (!diorama || typeof diorama !== 'object') {
- throw Object.assign(new Error('diorama is required — pass the object returned by compose_scene or get_scene.'), {
+ throw Object.assign(new Error('diorama is required: pass the object returned by compose_scene or get_scene.'), {
code: 'diorama_required',
});
}
diff --git a/packages/scene-mcp/src/tools/get-scene.js b/packages/scene-mcp/src/tools/get-scene.js
index ab2acbed0..2bbc54661 100644
--- a/packages/scene-mcp/src/tools/get-scene.js
+++ b/packages/scene-mcp/src/tools/get-scene.js
@@ -1,4 +1,4 @@
-// `get_scene` — fetch a saved diorama by id. Read-only.
+// `get_scene`: fetch a saved diorama by id. Read-only.
//
// Wraps GET /api/diorama?id=<uuid>.
diff --git a/packages/scene-mcp/src/tools/list-scenes.js b/packages/scene-mcp/src/tools/list-scenes.js
index c1f6bae99..5c49bb51f 100644
--- a/packages/scene-mcp/src/tools/list-scenes.js
+++ b/packages/scene-mcp/src/tools/list-scenes.js
@@ -1,4 +1,4 @@
-// `list_scenes` — browse recent or featured dioramas. Read-only.
+// `list_scenes`: browse recent or featured dioramas. Read-only.
//
// Wraps GET /api/diorama?list=recent|featured&limit=<n>.
@@ -11,7 +11,7 @@ export const def = {
title: 'List recent or featured dioramas',
annotations: { readOnlyHint: true, idempotentHint: false, openWorldHint: true },
description:
- 'Browse the public diorama gallery — the most recent saved worlds, or the featured set. Returns a list of cards (id, title, mood, thumbnail/preview, view count) plus each world\'s viewer URL. Read-only.',
+ 'Browse the public diorama gallery: the most recent saved worlds, or the featured set. Returns a list of cards (id, title, mood, thumbnail/preview, view count) plus each world\'s viewer URL. Read-only.',
inputSchema: {
list: z
.enum(['recent', 'featured'])
@@ -23,7 +23,7 @@ export const def = {
.min(1)
.max(50)
.optional()
- .describe('How many to return (1–50, default 24).'),
+ .describe('How many to return (1-50, default 24).'),
},
async handler(args) {
const scope = args?.list === 'featured' ? 'featured' : 'recent';
diff --git a/packages/scene-mcp/test/registration.test.mjs b/packages/scene-mcp/test/registration.test.mjs
index edd4efdd1..2968d282e 100644
--- a/packages/scene-mcp/test/registration.test.mjs
+++ b/packages/scene-mcp/test/registration.test.mjs
@@ -2,7 +2,7 @@
//
// Importing src/index.js is side-effect-free: the stdio transport only
// connects when the file is the process entry point, and buildServer() needs
-// no key or signer. These tests run offline — they never touch the network.
+// no key or signer. These tests run offline; they never touch the network.
//
// Run: node --test packages/scene-mcp/test/registration.test.mjs
@@ -41,7 +41,7 @@ test('non-read-only tools set destructiveHint explicitly (spec default is TRUE w
assert.equal(
typeof tool.annotations.destructiveHint,
'boolean',
- `${tool.name} is not read-only — destructiveHint must be explicit`,
+ `${tool.name} is not read-only; destructiveHint must be explicit`,
);
}
}
diff --git a/prompts/swarm-100/packages-batch-08.md b/prompts/swarm-100/packages-batch-08.md
deleted file mode 100644
index e30303925..000000000
--- a/prompts/swarm-100/packages-batch-08.md
+++ /dev/null
@@ -1,95 +0,0 @@
-# Package audit batch 08: pumpfun-skills, react, readme-3d, reputation, retarget, scene-mcp
-
-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
-
-For each package below: read its package.json and README.md, then find its
-tests (its own directory and tests/ at the repo root).
-
-## Task: audit and fix these packages
-
-Packages in this batch:
-
-- packages/pumpfun-skills
-- packages/react
-- packages/readme-3d
-- packages/reputation
-- packages/retarget
-- packages/scene-mcp
-
-For each one:
-
-1. README: present (repo standard is 100% coverage under packages/) and
- accurate: what it does, install and use, public exports, one runnable
- example. Run the example for real.
-2. Exports resolve: import the package's declared entry points with node and
- confirm they load without error.
-3. Tests: run the tests that cover it; if its core path has none, add a real
- one.
-4. Hard-rule scan across its source: no stubs, no mock fallbacks, no
- unfinished markers, no commented-out code. Fix at root.
-5. package.json sanity: name, version, license, exports and files fields
- consistent with what is actually shipped.
-
-## Definition of done
-
-- [ ] Every package in the batch: README verified accurate, example executed,
- entry points imported cleanly, tests green with core-path coverage.
-- [ ] Zero hard-rule violations remain in these packages.
-- [ ] npm run check:rules -- --paths <touched files> exits 0.
-
-## 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/packages-batch-08.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.