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: replace em-dashes and reword two prose comments the rules lint read as commented-out code in the approvals, strategy, domains and desktop runtime files

nirholas66624fbe184f
diff --git a/api/_lib/agent-strategy-runtime.js b/api/_lib/agent-strategy-runtime.js
index da1900c01..4b6aa600f 100644
--- a/api/_lib/agent-strategy-runtime.js
+++ b/api/_lib/agent-strategy-runtime.js
@@ -395,7 +395,7 @@ async function evaluateEntries({ equip, agent, launches, nowMs, maxEntries = 3 }
if (sinceMin < config.risk.cooldown_minutes) return results;
}
- // Concurrency — never exceed the strategy's max concurrent open positions. In
+ // Concurrency: never exceed the strategy's max concurrent open positions. In
// ask mode a pending approval holds a slot too, so the owner is never asked for
// more buys than the strategy may hold at once.
const [openCnt] = await sql`
diff --git a/api/_lib/approvals.js b/api/_lib/approvals.js
index 508b0e8b7..bad091aaf 100644
--- a/api/_lib/approvals.js
+++ b/api/_lib/approvals.js
@@ -9,7 +9,7 @@
// risk notes. Approving:
//
// 1. requires the payload hash the owner was shown, so a screen that went
-// stale (or a link someone edited) can never approve a different action;
+// stale (or a link someone edited) can never approve a different action, and
// 2. flips pending -> approved in ONE conditional UPDATE that also checks the
// expiry, so a double tap, a push tap racing a web click, or an approval
// after the deadline cannot execute twice or late;
diff --git a/api/_lib/cloud-domains.js b/api/_lib/cloud-domains.js
index 86064fd76..f5c284c04 100644
--- a/api/_lib/cloud-domains.js
+++ b/api/_lib/cloud-domains.js
@@ -2,7 +2,7 @@
// domain_* MCP tools and /api/domains.
//
// Authenticated as the Cloud Run runtime service account through
-// getGcpAccessToken (gcp-auth.js). Every call is a real request to the API;
+// getGcpAccessToken (gcp-auth.js). Every call is a real request to the API, and
// nothing here is cached or faked. Callers that want caching wrap these (see
// domains-service.js).
//
diff --git a/api/agents/agent-strategy-objects.js b/api/agents/agent-strategy-objects.js
index 1d614ba67..f3ca62e43 100644
--- a/api/agents/agent-strategy-objects.js
+++ b/api/agents/agent-strategy-objects.js
@@ -309,7 +309,7 @@ async function handleClose(req, res, id) {
return json(res, 200, { data: result.data });
}
-// POST mode — switch one equipped strategy between auto (buy within caps) and ask
+// POST mode: switch one equipped strategy between auto (buy within caps) and ask
// (file an approval for every buy). Going to auto arms autonomous buying, so it
// needs the real-funds agreement; going to ask never does.
async function handleMode(req, res, id) {
@@ -352,7 +352,7 @@ async function handleApprovals(req, res, id) {
return json(res, 200, { data: { approvals: rows } });
}
-// POST approve — the owner's explicit yes for one ask-mode buy. Executes the exact
+// POST approve: the owner's explicit yes for one ask-mode buy. Executes the exact
// stored payload once, through every spend guard, at the moment of approval.
async function handleApprove(req, res, id) {
const owned = await loadOwned(req, res, id);
diff --git a/api/audit-log.js b/api/audit-log.js
index 82d792de5..3336bce7b 100644
--- a/api/audit-log.js
+++ b/api/audit-log.js
@@ -1,6 +1,6 @@
// GET /api/audit-log — JSON list of the caller's audit events
// GET /api/audit-log?format=csv — CSV download (same data, last 90 days)
-// GET /api/audit-log?action_prefix=approval_ — only actions starting with the
+// GET /api/audit-log?action_prefix=approval_ : only actions starting with the
// prefix (the /approvals page's audit panel). Lowercase letters and _ only.
//
// Backs the "Action log" panel on /dashboard-next/account. Reads from the
diff --git a/api/strategies.js b/api/strategies.js
index 7a0805ef4..8ca3884d8 100644
--- a/api/strategies.js
+++ b/api/strategies.js
@@ -355,7 +355,7 @@ async function handlePublish(req, res, auth, id) {
return json(res, 200, { data: publicStrategy(updated, perf.get(id), names.get(auth.userId)) });
}
-// POST /api/strategies/preview { config, hours?, network? } — public and read-only:
+// POST /api/strategies/preview { config, hours?, network? }: public and read-only:
// replays recorded launches through the live entry filter and research gates.
async function handlePreview(req, res) {
const rl = await limits.strategyPreviewIp(clientIp(req));
diff --git a/apps/desktop/src/runtime/strategy-schema.js b/apps/desktop/src/runtime/strategy-schema.js
index aad7193bb..bdc8b9486 100644
--- a/apps/desktop/src/runtime/strategy-schema.js
+++ b/apps/desktop/src/runtime/strategy-schema.js
@@ -1,4 +1,4 @@
-// Strategy Object config — the validated, versioned rule set behind a Strategy.
+// Strategy Object config: the validated, versioned rule set behind a Strategy.
//
// A strategy is NOT free text: it is a structured plan with a real schema. This
// module is the single source of truth for that schema. It normalizes arbitrary
@@ -9,7 +9,7 @@
//
// Critically, the strategy's own caps (per-trade size, slippage, concurrency)
// are ADDITIONAL constraints layered on top of the agent's server-side spend
-// policy — never a way around it. The runtime sizes a buy from `sizing.amount_sol`
+// policy: never a way around it. The runtime sizes a buy from `sizing.amount_sol`
// but the trade still passes through the full guard + custody path, so a strategy
// can never exceed the spend leash. Ever.
//
@@ -81,7 +81,7 @@ export const STRATEGY_CONFIG_DEFAULTS = Object.freeze({
},
});
-// Hard ceilings the schema enforces regardless of input — defensive bounds so a
+// Hard ceilings the schema enforces regardless of input: defensive bounds so a
// stored config can never carry a nonsense number into the runtime.
const BOUNDS = Object.freeze({
amount_sol: { min: 0.0001, max: 100 },
@@ -199,7 +199,7 @@ export function normalizeStrategyConfig(raw) {
},
exits: {
take_profit_pct: numOrNull(x.take_profit_pct, BOUNDS.take_profit_pct),
- // stop_loss is mandatory and always present — default applied if absent/invalid.
+ // stop_loss is mandatory and always present: default applied if absent/invalid.
stop_loss_pct: clampNum(x.stop_loss_pct, d.exits.stop_loss_pct, BOUNDS.stop_loss_pct),
trailing_stop_pct: numOrNull(x.trailing_stop_pct, BOUNDS.trailing_stop_pct),
max_hold_minutes: intOrNull(x.max_hold_minutes, BOUNDS.max_hold_minutes),
@@ -272,7 +272,7 @@ export function validateStrategyConfig(raw) {
errors.push({ field: 'sizing.amount_sol', message: 'Per-trade size must be greater than 0 SOL.' });
}
if (!(config.exits.stop_loss_pct > 0)) {
- errors.push({ field: 'exits.stop_loss_pct', message: 'A stop-loss is required — every strategy must define its downside.' });
+ errors.push({ field: 'exits.stop_loss_pct', message: 'A stop-loss is required: every strategy must define its downside.' });
}
if (
config.entry.min_market_cap_usd != null &&
@@ -299,7 +299,7 @@ export function validateStrategyConfig(raw) {
/**
* Evaluate one real launch against a strategy's entry conditions. Pure +
- * synchronous. Returns { pass, reasons } — reasons always explains the verdict
+ * synchronous. Returns { pass, reasons }: reasons always explains the verdict
* (kept on rejections too, for the runtime's evaluation log).
*
* @param {object} config normalized strategy config
@@ -314,11 +314,11 @@ export function matchesEntry(config, launch, nowMs) {
if (!launch || !launch.mint) return { pass: false, reasons: ['no_mint'] };
- // Age gate — only act on genuinely recent launches.
+ // Age gate: only act on genuinely recent launches.
if (e.max_age_minutes != null && launch.created_at) {
const ageMin = (nowMs - Number(launch.created_at)) / 60000;
if (!Number.isFinite(ageMin) || ageMin < 0) {
- // Clock skew / bad timestamp — treat as fresh, don't reject.
+ // Clock skew / bad timestamp: treat as fresh, don't reject.
} else if (ageMin > e.max_age_minutes) {
return { pass: false, reasons: [`too_old:${Math.round(ageMin)}m`] };
} else {
@@ -326,7 +326,7 @@ export function matchesEntry(config, launch, nowMs) {
}
}
- // SOL-quote requirement — the agent wallet trades in SOL on this path.
+ // SOL-quote requirement: the agent wallet trades in SOL on this path.
if (e.require_sol_quote && launch.is_usdc_pair === true) {
return { pass: false, reasons: ['quote_not_sol'] };
}
@@ -472,7 +472,7 @@ export function evaluateResearchGates(config, report) {
/**
* Decide whether an open position should exit, given a live re-quote. Pure.
- * Returns { exit, reason } — reason ∈ take_profit|stop_loss|trailing_stop|timeout.
+ * Returns { exit, reason }: reason ∈ take_profit|stop_loss|trailing_stop|timeout.
*
* @param {object} config normalized config
* @param {object} pos { entry_lamports, peak_value_lamports, opened_at(ms) }
@@ -484,7 +484,7 @@ export function shouldExit(config, pos, currentValueLamports, nowMs) {
const entry = Number(pos.entry_lamports || 0);
const cur = Number(currentValueLamports || 0);
- // Time-based exit is independent of price — check it even with no entry basis.
+ // Time-based exit is independent of price: check it even with no entry basis.
if (x.max_hold_minutes != null && pos.opened_at) {
const heldMin = (nowMs - Number(pos.opened_at)) / 60000;
if (heldMin >= x.max_hold_minutes) return { exit: true, reason: 'timeout' };
diff --git a/public/push-sw.js b/public/push-sw.js
index e5ddce54d..0659fe458 100644
--- a/public/push-sw.js
+++ b/public/push-sw.js
@@ -9,7 +9,7 @@
// • notificationclick — focus or open the target URL (tagged ?source=push so
// the app records a `returned` funnel event), and beacon a
// `opened` event so sent→opened→returned is measurable.
-// • approvals — an approval push (api/_lib/approvals.js) gets Approve
+// • approvals : an approval push (api/_lib/approvals.js) gets Approve
// and Deny buttons. Tapping one posts the decision with the
// payload hash the notification showed, so the server only
// executes that exact action, then shows the outcome. A