Three pains every sales and BD team hits weekly. Each one is what your reps actually complain about, not what a feature page would call them.
A single L1 ran from 8 percent to 41 percent of the bag over six months, and the rebalance reminder lived in a Notion you stopped opening in March.
An LSD restaking position was paying 6.4 percent until the protocol shipped a contract upgrade, and you found out from a Twitter thread two days after the depeg.
Year-end tax pass starts in February because no one tracked which lots were sold at a loss, and the CPA bills another four thousand to reconstruct the chain of custody.
Every pipeline below is a shape you wire on the canvas using the crew and tools further down. Not a feature we ship for you, a pattern you configure.
PortfolioManager reads balances across EVM chains, Bitcoin, and Solana through scoped read-only RPC tools, scores single-asset, single-chain, and single-venue concentration, and flags any line above the band in your policy. Static context holds the policy so the score is consistent week over week.
ChiefRiskOfficer computes parametric and historical VaR plus CVaR, stress tests against March 2020 and the FTX cascade, and writes a one-page risk note. Hard drawdown thresholds are read from your policy, so the circuit breaker fires before the bag bleeds.
On a webhook from a watcher, the crew pulls the touching transaction, the counterparty contract source, and recent funding flows. The rag_retrieve tool fetches prior memos on that protocol so the alert lands with context, not just a hash. HITL gate before any rebalance memo is drafted.
ChiefRiskOfficer reads TVL, recent contract upgrades, oracle health, and audit dates on every active position, then scores residual yield against residual risk. Cross-run memory carries last week's protocol verdict so a fresh exploit headline drops the score immediately.
PortfolioManager pulls the full transfer history per wallet, models FIFO, HIFO, and Specific ID, and surfaces harvest candidates with a cited tx hash on every lot. ReportWriter assembles a one-page memo your CPA can sign off on. HITL gate before any lot is committed to the export.
QuantAnalyst pulls DXY, real yields, and global liquidity from FRED, ExecutionSpecialist scores rebalance cost in basis points, ReportWriter writes the regime call with cited series ids. Replay every chart back to its FRED series so next month's note is reproducible.
Real personas from the quant_firm crew. Each ships with a tuned system prompt and a default tool allowlist. Swap models per persona on the canvas.
Sizes positions with fractional Kelly, scores chain and venue concentration, and rebalances within the bands you set instead of chasing every market move.
Computes VaR and CVaR on the live portfolio, stress tests against 2008, March 2020, and FTX cascade, and enforces hard drawdown circuit breakers.
Tests parameter sensitivity on entry rules, exit rules, and rebalance bands. Reports signal quality with sample size, time period, and standard errors.
Hunts overfitting and regime fragility before you commit capital. Applies the Deflated Sharpe Ratio and demands walk-forward evidence on every thesis.
Decomposes rebalance cost into spread, market impact, and adverse selection in basis points so a DCA or TWAP plan does not silently eat your returns.
Distills the five specialist notes into a one-page decision document with a single ACTION line and confidence score. Useful for tax review and DAO updates.
Every tool below is a real shared tool from the Melaya bundle. Allowlist per agent; HITL-gate the writes; revoke any of them in one click.
Read native ETH and L2 balances, transaction receipts, and event logs through your own RPC. Read only, so the crew cannot move funds even if a prompt told it to.
Pull full token, NFT, and ERC-1155 transfer history per address across Etherscan-supported chains for cost basis and concentration scoring. Read only.
Read Bitcoin balances, UTXO sets, and fee estimates straight from mempool.space. The crew never holds an xpub, so cold storage stays cold.
Read SOL balances, SPL token positions, and signature history. Used by PortfolioManager when the bag spans Solana and EVM chains.
Cross-chain explorer reads for Bitcoin, Litecoin, Bitcoin Cash, and friends. Used when a hodler keeps long tail UTXO chains alongside the main bag.
Read DEX liquidity, pair depth, and trending pools so PortfolioManager flags positions where exit liquidity is thinner than the entry.
Read perp funding, OI, and liquidation cascades so the ChiefRiskOfficer stress test reflects current leverage in the system, not last cycle's.
Quote the bag at a consistent reference price for cost basis and reporting. Pairs with coinpaprika_tools as a fallback so a single API outage does not break a run.
Pull DXY, real yields, and global liquidity series so macro review carries a cited FRED id, not a vibe. Used by QuantAnalyst and ReportWriter.
General research, structured fetch for protocol audit pages, and the explicit HITL primitive. pause_for_human is the approval node in every rebalance and tax workflow.
Build the per-workflow store from your allocation policy, chain whitelist, prior tax filings, and DeFi position notes. Powers rag_retrieve and cross-run memory.
Every pipeline ships with three layers of knowledge access. Mix and match per agent on the canvas. No shared vector space with another tenant, no surprise reads, no opaque retrieval.
includeContextPer-pipeline documents appended to specific agents' input on every run. The ICP brief, playbook, pricing sheet, or won-deal email corpus. Whatever needs to be there before the agent thinks. You pick which personas get which docs.
rag_retrieveA scoped tool granted per-agent. When the agent decides it needs more depth, it queries the workflow's vector store on demand. Same knowledge base as Static context, accessed only when the model asks for it.
pipeline_memoryPipeline-level state that carries from one run to the next. Yesterday's research is in scope for today's follow-up. The crew remembers what it already prospected, what got approved, what was sent. The audit log is the second-order knowledge base.
Sandbox tier is free with no card. Join the waitlist and we will email you the moment a slot opens.
← Back to every use case