Kive Labs — Beta Agents

Agentic Agents

Beta Agent Overview

experimental agents before main launch

LP Neutral Agent — SUI delta-neutral

Dry-run · awaiting funding

Long SUI in a Cetus ±20% LP, delta-hedged by a Bluefin SUI-PERP short. Goal: earn LP fees while staying market-neutral on SUI.

Strategy total

—

Since start

—

Weekly

—

LP + Hedge

—

Net SUI exposure

—

Hedge uPnL

—

🟢 LP earning leg

—

SUI in LP (on-chain)—

SUI in LP (sim)—

USDSUI in LP—

SUI price—

LP range—

Status—

Avg entry / deposit—

LP earned (fees + rewards)—

🟠 Hedge leg · Bluefin SUI-PERP

—

Short size—

Notional—

Avg entry price—

Mark price—

Unrealized PnL—

Funding earned—

Liq. price—

Leverage—

LP vs short—

⚡ Perp order · maker-chase

—

Pending action

—

Target short

—

Net exposure

—

Resting orders

—

Side	Type	Price	Size	Filled	Status	Age
waiting for data…

—

Idle assets (not deployed): —

waiting for data…

How it works & addresses ▸

ferp — Funding-Rate Arbitrage

Dry-run · awaiting funding

Network

SUI mainnet

Token

USDC (native)

Functions

Send + Receive

Per-send cap

10 USDC

Daily cap

50 USDC

Mode

DRY_RUN (safe)

Receive address

0x4cc59fd32d167ab06d98406b5a2cb0e28fc245415c1a069154ccebfdcc7b388b ⧉ copy

ferp — Funding-Rate Arbitrage

Dry-run · awaiting funding

Network

SUI mainnet

Token

USDC (native)

Functions

Send + Receive

Per-send cap

10 USDC

Daily cap

50 USDC

Mode

DRY_RUN (safe)

Receive address

0x4cc59fd32d167ab06d98406b5a2cb0e28fc245415c1a069154ccebfdcc7b388b ⧉ copy

ferp chase — paper

ferp's live chase gave up on 60 of 91 attempts (66%); the worst run burned 3.5 h over 15 consecutive failures. This replays the candidate rules against the live book and times each fill. No orders, no keys.

Read the fill model before trusting a number. A sell is counted filled when the best bid reaches its price. That ignores queue position, so absolute times are optimistic. The comparison between rules is what this is for. If paper fills in seconds while live does not, the bottleneck is queue and size — not price movement.

Across all sessions

Rule	Fill rate	Median fill	Worst fill	Avg reposts	Paid spread
waiting for data…

Most recent sessions

When	Market	Side	Rule	Fill	Reposts	Bid moved
waiting for data…

mock — wallet watcher

Read-only · no keys

Bitkub’s published proof-of-reserve wallets: 101 on Bitcoin and 3 on Sui. Addresses come from CoinMarketCap, which shows them truncated — these are the full values, each checked against the displayed prefix and suffix. Every balance is read from the chain, not taken from CMC, whose figures were wrong elsewhere: it reported 29.8M XRP in a wallet holding nothing.

Total · BTC + ETH

—

excludes ERC-20 tokens

Bitcoin wallet

—

—

Sui wallets

—

—

Verification

—

checksums + node agreement

—

period →

—

Current holdings — amount and value On date live

	Asset	Address	Amount	Price	Value	Share
waiting for data…

Wallet 1

Bitcoin

P2WSH bech32 valid —

bc1qdvt3jnrfxnc8cy30ep9r9p9vwxjf7l66ylz63zam4883ntm9zxgqfzptar ⧉ copy

Balance

—

—

Vs cost basis

—

—

Average entry

—

—

Activity

—

—

Transaction	Net BTC	USD then	USD now	Fee	Block	Date
waiting for data…

#	Address	SUI	Value USD	Share
waiting for data…

5 min bid 45-45

checkingdry-run

Live Polymarket agent for BTC 5-minute Up/Down prediction rounds. Current bot rests BUY 5 Up @ 0.40 once per 5m window while the kill-switch is off; the dashboard tracks Magic/API wallet balance, open orders, positions, recent fills, and trading P&L. Private keys and API secrets are never shown here.

USDC balance

—

Positions value

—

Marked P&L

—

Open orders

—

Trading P&L since start · daily

—

5 min bid 45-45 · Paper mock

—

Equity

—

Mock P&L

—

Matched / missed

—

Fill rate

—

Up / Down ask

—

Per strategy

Long · Up @0.45

P&L — · Eq —

matched/miss — · fill —

Short · Down @0.45

P&L — · Eq —

matched/miss — · fill —

Long mock P&L · Up @0.45

Short mock P&L · Down @0.45

—

Long · Up @0.45 — order history

Time	Status	Best ask	Result	P&L	Short · Down

Short · Down @0.45 — order history

Time	Status	Best ask	Result	P&L	Long · Up

Fake-money simulation only. Posts no real order. It samples public Polymarket best asks; Up and Down paper orders each match only if ask <= 0.45.

Track record · BTC 5m rounds

Realized

—

Unrealized

—

Total P&L

—

Record

—

Win rate

—

Time	Window	Entry	Result	Payout	P&L

💰 Funding · deposit USDC on Polygon to this address

—

USDC balance—

Token approvals—

Recent Up 49 paper orders

Market	Side	Shares	Avg	Now	uP&L

Open orders

Market	Side	Outcome	Price	Size

Safety · Max order — · Market orders — · Mode — · Chain —

updated —

BTC 5m Bid V2

draftfake only

Fresh V2 of the BTC 5-minute Up/Down paper-bid mechanism. It is designed to work like V1/45-45 — sample real public order books, paper bid both sides, and count only realistic fills — but it starts with no copied V1 data.

Mode

Paper

Data state

Blank

Mechanism

V1-like

Real money

V2 mechanism

1. Read markets

Use public Polymarket BTC 5-minute Up/Down order-book data.

2. Paper bid

Create fake bids for both Up and Down sides, same realism rule as V1.

3. Track fills

Only mark a fake order matched if the public ask makes it executable.

Data separation

V1 data copied?	No
Current V2 runner	Not created yet
Planned files	apm_v2_paper.json, apm_v2_pnl.json, apm_v2_leg_pnl.json

V2 is a clean beta-agent shell. No V1 data/state was copied, no wallet is attached, and no real orders can be placed.

5 min bid 49-49

paperfake only

Fake-money Polymarket strategy for BTC 5-minute Up/Down rounds. It paper bids BUY 5 Up @0.49 and BUY 5 Down @0.49 every 5-minute window using real public order-book data. No wallet, no real orders, no real money.

Mode

No wallet

Fake equity

—

Mock P&L

—

Matched / missed

—

5 min bid 49-49 · Paper mock

—

Equity

—

Mock P&L

—

Matched / missed

—

Fill rate

—

Up / Down ask

—

Per strategy

Long · Up @0.49

P&L — · Eq —

matched/miss — · fill —

Short · Down @0.49

P&L — · Eq —

matched/miss — · fill —

Long mock P&L · Up @0.49

Short mock P&L · Down @0.49

—

Long · Up @0.49 — order history

Time	Status	Best ask	Result	P&L	Short · Down

Short · Down @0.49 — order history

Time	Status	Best ask	Result	P&L	Long · Up

Fake-money simulation only. Posts no real order. It samples public Polymarket best asks; Up and Down paper orders each match only if ask <= 0.49.

updated —

Covered Call Agent

draftfake only

Covered-call style beta agent: hold BTC and sell upside exposure to collect income. This page is a strategy shell only for now — no API key, no wallet, no live orders, no real money.

Mode

Paper

Underlying

BTC

Income source

Call premium

Risk

Capped upside

Strategy idea

1. Hold BTC

Keep spot exposure so the agent benefits if BTC rises up to the target.

2. Sell upside

Sell a call-like contract above current price to collect premium.

3. Manage risk

If BTC runs above strike, upside is capped; if BTC falls, premium only cushions part of the drawdown.

Next wiring

Market data	Binance BTC options chain: bid/ask, mark, IV, Greeks
Decision	Pick expiry + strike based on target yield and delta
Execution	Start paper-only; live trading needs explicit approval

Covered Call Agent is draft-only. It does not post orders or touch funds.