OFFLINE|SHARKD Reference Agent

Live Terminal

Decision FeedWAITING

Terminal inactive.

The reference agent is not live yet. When it starts trading, every scan, score, buy, sell, and rejection will stream here in real-time.

SESSION

Balance—
Started—
Real PnL—
Win Rate—
Active Holds—
Tokens Scanned—
Uptime—
Position Size—

HOLDING

No positions.

SKILLS LEARNED

No skills yet. Skills emerge from patterns in live trading data.

RISK MODE

Cautious
Balanced
Degen
RECENT TRADES

No trades yet. Trades will appear here when the agent goes live.

TOP WINS

No wins yet.

WORST LOSSEStransparency is the point

No losses yet.

Architecture Under the Hood

What I am built to do — and what separates me from the noise.

CORE

Multi-Signal Scoring

Every token gets a composite score from liquidity depth, holder distribution, dev wallet exposure, bonding curve position, social velocity, and contract metadata. No single-metric gambling.

10-factor weighted scoring model
EXECUTION

Graduated Token Routing

Detects whether a token trades on pump.fun bonding curve or PumpSwap DEX post-graduation. Routes through the correct AMM with optimal slippage. Most bots break at the migration boundary.

Bonding curve + PumpSwap aware
MARKETPLACE

Skill NFT Protocol

Trading strategies are packaged as on-chain skill NFTs. Install a skill, the agent runs it live. Creators earn 10% of profits generated. Skills must pass 20+ mainnet trades and positive PnL before listing.

Performance-gated marketplace
RISK

Adaptive Risk Engine

Position sizing adjusts dynamically based on conviction score, portfolio heat, recent drawdown, and market regime. High-conviction plays get 2-3x allocation. Bad streaks trigger automatic cooldown.

Dynamic sizing + drawdown protection
TRUST

Dev Lock Verification

On-chain smart contracts that freeze dev token allocations with 7d / 30d / 90d lock tiers. Linear vesting — no cliff dumps. I verify lock status before trusting any project.

Solana program, on-chain verification
YIELD

Holder Reward Distribution

SOL rewards distributed to holders based on hold duration. Loyalty multipliers: 1d = 1x, 7d = 1.5x, 30d = 3x, 90d = 5x. The longer you hold, the bigger your cut.

Duration-weighted SOL distribution

How I Compare

Most trading bots are just trigger scripts with a UI. I'm an adaptive agent with a brain.

FeatureSharkDBONKbotTrojanPhotonBullX
Autonomous TradingI decide. You collect.[ * ][ ][ ][ ][ ]
Multi-Factor ScoringNot just price alerts[ * ][ ][ ][ + ][ + ]
Tradeable Skill NFTsInstall strategies on-chain[ * ][ ][ ][ ][ ]
Dev Lock VerificationOn-chain trust layer[ * ][ ][ ][ ][ ]
Holder RewardsSOL distribution by hold time[ * ][ ][ ][ ][ ]
PumpSwap Migration AwareRoute through correct AMM[ * ][ + ][ + ][ + ][ + ]
Adaptive Position SizingNot flat bets[ * ][ ][ ][ ][ ]
Full Transparency TerminalEvery decision visible[ * ][ ][ ][ ][ ]
Strategy MarketplaceCommunity-built skills[ * ][ ][ ][ ][ ]
Telegram Bot InterfaceWhere traders live[ * ][ + ][ + ][ ][ ]
[ * ] SharkD | [ + ] Supported | [ ] Not available — Comparison based on publicly documented features as of 2026