Skip to main content
Glama

FarmDash OpenClaw Skills

Public FarmDash MCP server

This repository includes the runnable MIT-licensed FarmDash stdio MCP adapter in mcp-server/. It exposes the same 94-tool contract published at https://www.farmdash.one/.well-known/mcp.json and calls the live FarmDash API without receiving customer private keys.

Build and verify from a clean checkout:

git clone https://github.com/Parmasanandgarlic/farmdash-openclaw-skills.git
cd farmdash-openclaw-skills
npm ci --prefix mcp-server
npm run build --prefix mcp-server
npm run smoke --prefix mcp-server
node mcp-server/dist/index.js

For MCP clients, configure node with the absolute path to mcp-server/dist/index.js. FARMDASH_API_KEY is optional for Scout. The package metadata uses @farmdash/mcp-server, but do not use an npm install command until registry publication is independently verified. The public source checkout is the current installable surface; no published container image exists for this project.

10 open agent skills for DeFi intelligence on FarmDash — zero-custody and fail-closed. Skills are read-only unless their documented workflow explicitly reaches a user-authorized preparation/execution boundary. FarmDash skills never ask for private keys or seed phrases.

Skill

What it does

farmdash-trail-intelligence

Rank DeFi protocols, airdrops, points programs; Trail Heat + FarmScore inputs; sybil-policy risk

farmdash-signal-architect

Swap planning with explicitly user-signed payloads (EIP-191/EIP-712)

farmdash-futures-strategist

Perps planning with explicitly user-signed payloads

farmdash-hedge-warden

Hedge design and risk guards

farmdash-ledger-keeper

Portfolio books and position tracking

farmdash-supply-master

Supply / liquidity stewardship

farmdash-trail-marshal

Rotation planning across trails

farmdash-wagon-steward

Wagon (portfolio convoy) stewardship

farmdash-camp-guard

Safety policy enforcement

farmdash-autonomous-operator

Supervised operator routines (human-gated)

FARMDASH_AGENT_OPERATING_MODEL.md is the shared operating contract all ten skills run under.

Related MCP server: defi-yield-scanner-mcp

Install

Copy any skill directory into your agent's skills path (e.g. OpenClaw / Claude Code skills), or point your harness at this repo. Each skill is self-contained: SKILL.md (agent manual) + SKILL.json (tool manifest).

API access

  • Scout tier: free, keyless (fd_scout_free), 30 req/24h per IP. The MCP adapter is forward-compatible with a bounded onboarding grant if the live FarmDash onboarding response advertises one; clients must treat /api/v1/agent/status and /api/v1/agent/onboard as runtime truth rather than assuming the grant is deployed.

  • Pioneer/Syndicate: Bearer FARMDASH_API_KEY for higher limits and deep tools.

  • MCP manifest: https://www.farmdash.one/.well-known/mcp.json (v5.0.0, 94 runtime tools)

  • OpenAPI: https://www.farmdash.one/agents/openapi.yaml (54 primary operations / 75 total across 4 documents)

Commercial disclosure: skills may surface farmdash.one/go/* partner routes with mandatory disclosure; never on avoid-verdicts. Details: https://www.farmdash.one/fees.

License

MIT — see LICENSE.

Available Tools

94 tools
agent_onboardA
Read-onlyIdempotent

[Capability: available; tier: scout; API: /api/v1/agent/status; authority: /api/v1/agent/status] First-run contract: fetch live feature readiness, tier gates, error semantics, execution-mode selection (USER_SIGNED default vs opt-in BOUNDED_AUTONOMOUS), and the onboarding capability map before selecting tools. Before entering any wallet-changing execution workflow, query current execution authority with get_execution_authority for the active persisted FarmDash session: if an active valid bounded delegation already exists for that session, report its policy bounds instead of re-prompting; a different/new agentSessionId does not inherit another session’s grant and remains USER_SIGNED (Standard FarmDash), and you must explicitly obtain user choice/consent before requesting a one-time scoped delegation for BOUNDED_AUTONOMOUS (Autonomous FarmDash, EOA_V1_OWNER_LOCAL_SIGNER where sessionPublicKey == ownerAddress). This never enables execution. Takes no parameters.

Inputs: no named parameters.

analyze_futures_strategyB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus_or_x402; API: /api/v1/agent/futures/analyze-strategy; authority: /api/v1/agent/status] Primary research tool. Returns strategy object with confidence, regime, and simulation. For a delegated Hyperliquid API wallet, set agentAddress to the signer and accountAddress to the master/subaccount whose equity must size the trade; the venue relationship is verified and fails closed.

Inputs: coin (required), agentAddress (required), grantId, sessionId, +6 more.

analyze_protocol_riskB
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/protocol-risk; authority: /api/v1/agent/status] Deep risk analysis for a DeFi protocol. Pulls live signals from DeFiLlama (TVL, momentum, chain diversification, historical trends) and combines with Trail Heat scoring to produce a composite risk grade (low/moderate/elevated/high/critical) with actionable signals.

Inputs: protocolId (required), includeHistorical.

audit_allowance_riskB
Read-onlyIdempotent

[Capability: available; tier: scout_limited_pioneer_full; API: /api/v1/agent/camp-guard; authority: /api/v1/agent/status] Audit token approvals for unlimited, unknown, or high-dollar allowance risk.

Inputs: allowances, walletAddress.

audit_sybil_riskB
Read-onlyIdempotent

[Capability: available; tier: pioneer_plus; mock_supported: false; API: /api/v1/agent/sybil-audit; authority: /api/v1/agent/status] Research: audit 1-10 EVM addresses for sybil risk. Returns cluster risk score and hygiene suggestions.

Inputs: addresses (required).

authorize_virtuals_tender_v2B

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Submit an exact V2 EIP-712 signature produced by the customer ACP wallet. FarmDash first verifies deployment readiness and all committed providers; it still cannot create or fund jobs.

Inputs: nonce (required), tenderId (required), expiresAt (required), sessionId (required), +4 more.

autopilot_cycleB

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Run one authenticated autopilot cycle. Returns recommended actions.

Inputs: sessionId (required), agentAddress (required), sessionToken (required), accountState.

bind_virtuals_tender_jobB
Idempotent

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Report a customer-created on-chain ACP job ID for one committed role. FarmDash independently verifies the Base client, provider, and evaluator before binding.

Inputs: role (required), tenderId (required), sessionId (required), agentAddress (required), +2 more.

calculate_position_sizeB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus_or_x402; API: /api/v1/agent/futures/position-sizing; authority: /api/v1/agent/status] Translate risk constraints into size and leverage for a perp setup.

Inputs: equity (required), stopPrice (required), entryPrice (required), riskPercent, +2 more.

cancel_perp_orderA
Destructive

[Capability: available_with_tier_gate; tier: scout_capacity_signature_gated; API: /api/v1/agent/futures/cancel-order; authority: /api/v1/agent/status] USER_SIGNED requires fresh human approval; BOUNDED_AUTONOMOUS requires futures.cancel scope, a valid owner-signed policy and authenticated session, but not unused entry budget. Local venue signatures are required in both modes. Cancel stale or resting Hyperliquid orders with a venue-signed expiresAfter and FarmDash intent hash. For a delegated API wallet, accountAddress identifies the verified equity owner and is included as the signed vault/subaccount routing address when required. accepted is an application acknowledgement, not final venue cancellation: reconcile every requested ID through /api/v1/agent/futures/reconcile-order before dependent actions. Mixed errors are partial rejection; malformed or timeout outcomes are unknown. Venue nonce authority is network + signer + nonce across account/vault contexts; durable replay control fails closed before submission.

Inputs: coin (required), nonce (required), orderIds (required), expiresAt (required), +9 more.

cancel_virtuals_tenderB
Destructive

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Cancel an authenticated non-spendable ACP draft. It cannot erase or refund an already-created on-chain job.

Inputs: tenderId (required), sessionId (required), agentAddress (required), sessionToken (required).

compare_yieldsB
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/supply; authority: /api/v1/agent/status] Research: compare DeFi yield pools by APY, TVL depth, stablecoin exposure, IL risk, and sustainability score.

Inputs: limit, assets, chains, projects, +3 more.

configure_autopilotC
Idempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Configure autopilot strategies, assets, and risk parameters.

Inputs: assets (required), sessionId (required), strategies (required), agentAddress (required), +5 more.

confirm_autonomous_swapB
Idempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Submit on-chain transaction hash for independent RPC receipt verification and economic reconciliation.

Inputs: txHash (required), reservationId (required), agentSessionId (required), status, +4 more.

confirm_executionB
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/execute; authority: /api/v1/agent/status] Unavailable for generic public confirmation: FarmDash will not record a caller-supplied receipt. A trusted adapter and server-side transaction verifier must be deployed before this action can be enabled; use get_receipt only for verified receipt observation.

Inputs: txHash (required), intentId (required), details, sessionId, +2 more.

confirm_swapA
Idempotent

[Capability: server_verified; tier: session_bound; API: /api/agents/confirm; authority: /api/v1/agent/status] Confirm a compatibility swap only through FarmDash server verification of the owning session, canonical-chain receipt, exact committed fee asset/payer/recipient/amount, and required finality. A caller-supplied tx hash alone is not confirmation.

Inputs: txHash (required), sessionId (required), feeEventId (required), agentAddress (required), +1 more.

create_intentA
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/create; authority: /api/v1/agent/status] Plan: create a durable FarmDashIntent. This records what an agent wants to do; it never prepares, signs, broadcasts, or executes.

Inputs: intent (required), sessionId, agentAddress, sessionToken.

create_mee_intentB
Idempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/mee; authority: /api/v1/agent/status] Create a preparation-only MEE-shaped cross-chain intent bundle. Creation is not authoritative simulation, submission, or settlement.

Inputs: steps (required), ownerAddress (required).

create_sessionB
Idempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/session; authority: /api/v1/agent/status] Create a persistent agent session. Normal callers must include a fresh EIP-712 or EIP-191 session signature. Returns a one-time sessionToken.

Inputs: agentAddress (required), nonce, issuedAt, agentName, +2 more.

evaluate_virtuals_tenderB
DestructiveIdempotent

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Request evaluation or resume settlement for a fully bound customer-funded ACP tender. FarmDash can only complete/reject jobs committed to its evaluator identity.

Inputs: tenderId (required), sessionId (required), agentAddress (required), sessionToken (required).

execute_approved_intentB
Destructive

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/execute; authority: /api/v1/agent/status] Unavailable for generic public execution: the lifecycle API rejects client-supplied receipts until a trusted adapter and server-side transaction verifier are deployed. Use this only to observe the structured execution_verifier_unconfigured response; raw arbitrary calldata is rejected.

Inputs: intentId (required), error, status, txHash, +7 more.

execute_cycle_actionsC
DestructiveIdempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Inspect the cycle execution gate. Server-side ERC-4337 action submission is preview-only and returns autonomous_execution_preview_only until real calldata, session signatures, and bundler verification ship.

Inputs: jobId (required), sessionId (required), agentAddress (required), sessionToken (required).

execute_perp_orderB
Destructive

[Capability: available_with_tier_gate; tier: scout_capacity_signature_gated; API: /api/v1/agent/futures/execute-order; authority: /api/v1/agent/status] USER_SIGNED requires fresh human approval. BOUNDED_AUTONOMOUS requires a valid owner-signed futures policy, authenticated session, and a local API-wallet signature on every action; rollout remains preparation_only. Compatibility: submit a pre-signed EIP-712 Hyperliquid order after a 60-second parameter-bound research gate. market_ioc is a signed, price-protected IOC limit order with a live 100 bps adverse-price cap, never an unbounded market order; its outcome distinguishes full, partial, and zero fill, and an IOC resting response is unknown/inconsistent. For delegated API wallets, set accountAddress to the venue-verified equity owner used during analysis. FarmDash recovers the exact L1 signer, rejects an agentAddress mismatch, and preflights the owner maxBuilderFee. The signed action must disclose builder f=1 (0.1 bp/0.001% of filled notional), and expiresAt is venue-signed as expiresAfter. Venue nonce authority is network + signer + nonce across account/vault contexts; durable replay control fails closed before submission.

Inputs: coin (required), size (required), isBuy (required), nonce (required), +15 more.

execute_swapB
Destructive

[Capability: available_with_tier_gate; tier: scout_quote_signature_and_simulation_gated; API: /api/agents/swap; authority: /api/v1/agent/status] Compatibility route: validate a signed token-swap request after a fresh successful simulation and return transaction calldata. FarmDash does not broadcast the returned transaction. Generic execute_approved_intent remains unavailable.

Inputs: nonce (required), toToken (required), fromToken (required), signature (required), +9 more.

find_capital_routeA
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_quote_signature_and_simulation_gated; API: /api/agents/quote; authority: /api/v1/agent/status] Find the most efficient route between tokens/chains. Considers fees, slippage, and gas. Returns a provider-neutral market estimate (never executable, never attributed) — a wallet alone never triggers provider quoting. For execution, create a firm single-provider quote via the quote-intent API (LI.FI or Relay; 0x remains operator-paused), then simulate before preparing.

Inputs: toToken (required), fromToken (required), toChainId (required), fromAmount (required), +2 more.

get_agent_activityB
Read-onlyIdempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/receipts; authority: /api/v1/agent/status] Observe: list durable FarmDash execution receipts. Filter by intent_id and receipt status to review recent agent activity.

Inputs: limit, status, intentId, sessionId, +2 more.

get_agent_eventsA
Read-onlyIdempotent

[Capability: available; tier: pioneer_plus; mock_supported: false; API: /api/v1/agent/events; authority: /api/v1/agent/status] Poll Pioneer+ tenant-bound agent events using the owning authenticated session (sessionId, agentAddress, and sessionToken); no Scout or mock preview. Returns recent event records, not a public real-time stream.

Inputs: sessionId, agentAddress, sessionToken.

get_agent_performanceA
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus_or_x402; API: /api/v1/agent/performance; authority: /api/v1/agent/status] Review FarmDash fee-event activity, fees, protocol diversity, and reputation. This endpoint does not return fills, trade outcomes, realized PnL, win rate, or slippage.

Inputs: agentAddress (required).

get_approval_statusB
Read-onlyIdempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}; authority: /api/v1/agent/status] Approve: inspect whether an intent is still awaiting approval, has been approved, or was rejected.

Inputs: intentId (required), sessionId, agentAddress, sessionToken.

get_chain_breakdownB
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/chain-breakdown; authority: /api/v1/agent/status] Get protocol distribution aggregated by blockchain network. Returns counts, airdrops, and category coverage. Takes no parameters.

Inputs: no named parameters.

get_cycle_statusB
Read-onlyIdempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Poll the status of an asynchronous autopilot cycle job.

Inputs: jobId (required), sessionId (required), agentAddress (required), sessionToken (required).

get_event_stream_snapshotC
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/stream; authority: /api/v1/agent/status] Observe: read a JSON snapshot from the agent event stream for a live session.

Inputs: sessionId (required), agentAddress (required), sessionToken (required), limit, +2 more.

get_execution_authorityC
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Query execution authority configuration (USER_SIGNED vs BOUNDED_AUTONOMOUS), zero-custody confirmation, and active delegation status.

Inputs: grantId, sessionId, ownerAddress, sessionToken, +1 more.

get_farming_contextC
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/context; authority: /api/v1/agent/status] Read the shared FarmingContext for a live agent session.

Inputs: sessionId (required), agentAddress (required), sessionToken (required).

get_futures_accountA
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus_or_x402; API: /api/v1/agent/futures/account-state; authority: /api/v1/agent/status] Inspect Hyperliquid equity, margin, positions, and drawdown state. agentAddress identifies the API signer; accountAddress identifies the master/subaccount that owns the equity (omit only for a direct signer). Hyperliquid userRole must verify delegation. Daily/weekly values are conservative venue-derived loss-pressure metrics, not full period returns; history ambiguity fails closed.

Inputs: agentAddress (required), accountAddress.

get_historical_trailheatB
Read-onlyIdempotent

[Capability: available; tier: pioneer_plus; mock_supported: false; API: /api/v1/agent/historical-trailheat; authority: /api/v1/agent/status] Query historical Trail Heat snapshots, 1-365 days back. Returns score trends and momentum flags.

Inputs: days, protocolId.

get_idle_capitalC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/balances; authority: /api/v1/agent/status] Locate idle stablecoins or native assets across all supported chains.

Inputs: address (required).

get_mee_intent_statusA
Read-onlyIdempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/mee; authority: /api/v1/agent/status] Read the preparation status of a MEE-shaped intent. Created or simulated state is not proof of submission or settlement.

Inputs: id (required).

get_payments_bootstrapA
Read-onlyIdempotent

[Capability: available; tier: scout; API: /api/v1/payments/bootstrap; authority: /api/v1/agent/status] Free, provider-free payment bootstrap terms: how an agent holding only native Base ETH obtains metered access, including native-ETH payment intents, FarmDash prepaid credit, and the single-use execution-bootstrap grant. This tool is read-only. It contacts no swap provider, performs no mutation, and grants no execution authority. Issuing an intent or a grant requires an explicit authenticated POST with a server-issued, purpose-specific signature challenge, which this tool does not perform.

Inputs: no named parameters.

get_portfolio_summaryC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/balances; authority: /api/v1/agent/status] Get high-level portfolio metrics: total USD value, chain distribution, and capital efficiency score.

Inputs: address (required).

get_position_healthB
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/performance; authority: /api/v1/agent/status] Compatibility alias for agent activity/performance metrics. Current endpoint does not return DeFi position APR, PnL, impermanent loss, or rebalance health.

Inputs: address (required).

get_protocol_metadataA
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/protocols; authority: /api/v1/agent/status] Research: get catalog metadata for a specific DeFi protocol — status, chains, description, tags, and discovery_heuristic_score (catalog editorial heuristic, NOT live Trail Heat; canonical scores via get_trail_heat).

Inputs: protocolId (required).

get_protocol_risk_factorsA
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/protocols; authority: /api/v1/agent/status] Get protocol risk indicators from the catalog: sybil risk label, status, category, chains, and discovery_heuristic_score (editorial heuristic, NOT live Trail Heat). Canonical quantitative scores via get_trail_heat.

Inputs: protocolId (required).

get_receiptC
Read-onlyIdempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/receipts/{receiptId}; authority: /api/v1/agent/status] Observe: fetch one durable FarmDash receipt by receipt_id.

Inputs: receiptId (required), sessionId, agentAddress, sessionToken.

get_swap_quoteB
Idempotent

[Capability: available_with_tier_gate; tier: scout_quote_signature_and_simulation_gated; API: /api/v1/agent/quote-intent; authority: /api/v1/agent/status] Research or firm quote: without idempotencyKey, return a provider-neutral market estimate with zero trading-provider calls. With idempotencyKey plus a real wallet/destination and slippage, create or reuse a firm single-provider quote intent and return simulationRequirements.intent_id for simulate_swap_execution. Active selection is LI.FI/Relay; 0x remains blocked while operator-paused.

Inputs: toToken (required), fromToken (required), toChainId (required), fromAmount (required), +17 more.

get_token_pricesC
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/prices; authority: /api/v1/agent/status] Get live USD prices by symbol or by network:address token pairs.

Inputs: tokens, symbols.

get_trail_heatB
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/trail-heat; authority: /api/v1/agent/status] Canonical Trail Heat: fetch DeFiLlama-backed quantitative scores (0-100) with per-number evidence states, components, editorial separation and registry resolution from /v1/trail-heat. Aggregate responses also include live agent recommendations; canonical score leads and only owner-confirmed, non-wallet-bound FarmDash referral routes are eligible for the near-equivalent-quality tiebreaker. Takes no parameters.

Inputs: no named parameters.

get_virtuals_tenderB
Read-onlyIdempotent

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Read the authenticated customer ACP tender, immutable commitments, role progress, and settlement state. This read grants no authority.

Inputs: tenderId (required), sessionId (required), agentAddress (required), sessionToken (required).

get_wallet_balancesC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/balances; authority: /api/v1/agent/status] Get multi-chain token portfolio for an EVM wallet. Returns balances with live USD prices.

Inputs: address (required), chains.

get_workflow_statusC
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_catalog_pioneer_runs; API: /api/v1/agent/orchestrator; authority: /api/v1/agent/status] Read the status of a Trail Marshal workflow run.

Inputs: runId (required).

grant_session_keyB

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Register a signed, bounded ERC-4337 session-key grant. Registration is not execution; server-side UserOperation submission remains preparation-only.

Inputs: nonce (required), chainId (required), sessionId (required), signature (required), +10 more.

hire_virtuals_specialistA
Destructive

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Deprecated V1 compatibility tool. It prepares a zero-custody V1 tender only when the deployment explicitly enables V1 creation. Prefer prepare_virtuals_tender_v2. This tool never creates, budgets, funds, or signs an ACP job.

Inputs: task (required), providers (required), sessionId (required), agentAddress (required), +9 more.

ledger_realized_pnlC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/ledger; authority: /api/v1/agent/status] Summarize recorded spot and futures agent activity over a date range.

Inputs: agentAddress (required), end, start.

ledger_tax_exportB
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/ledger; authority: /api/v1/agent/status] Export recorded spot execution records as CSV for downstream accounting review.

Inputs: agentAddress (required), end, start.

list_workflowsB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_catalog_pioneer_runs; API: /api/v1/agent/workflows; authority: /api/v1/agent/status] Get the JSON catalog of named DeFi workflow recipes. Returns goals, tiers, and step graphs.

Inputs: filter.

optimize_portfolioC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/optimize-portfolio; authority: /api/v1/agent/status] Get unranked catalog research candidates. Returns WAIT analysis-only: exact pool identity, wallet eligibility, policy, costs and stay-put comparison remain required. Campaign confirmation does not establish safety.

Inputs: riskPreference (required), currentProtocols (required), budget, chains.

patch_farming_contextB
Idempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/context; authority: /api/v1/agent/status] Patch the shared FarmingContext for a live agent session. Server controls revision, sessionId, and agentAddress.

Inputs: patch (required), sessionId (required), agentAddress (required), sessionToken (required).

pause_autopilotB
Destructive

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Pause the backend cycle scheduler for an active agent session. This does not close positions, revoke venue authority, or prove chain-state changes.

Inputs: sessionId (required), agentAddress (required), sessionToken (required).

plan_workflowB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_catalog_pioneer_runs; API: /api/v1/agent/orchestrator; authority: /api/v1/agent/status] Build a Trail Marshal quality gate for a named workflow. Does not execute steps.

Inputs: workflowId (required), agentAddress, installedSkills.

policy_check_intentB
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/policy-check; authority: /api/v1/agent/status] Readiness probe for the generic intent policy gate. Public callers cannot provide trusted policy decisions; until the server policy evaluator is configured this returns policy_evaluator_unconfigured. Tool discovery does not grant policy authority.

Inputs: intentId (required), policy, context, sessionId, +2 more.

prepare_autonomous_swapA

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Prepare an autonomous spot swap within signed delegation bounds from authoritative quote and simulation records. FarmDash verifies 13 policy invariants, locks an atomic budget reservation, and computes the locked calldata digest. The agent signs and broadcasts in its local runtime.

Inputs: intentId (required), ownerAddress (required), simulationId (required), agentSessionId (required), +9 more.

prepare_intentA
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/prepare; authority: /api/v1/agent/status] Execute: validate adapter support and prepare an intent after policy, simulation, and approval gates pass. This does not broadcast a transaction.

Inputs: intentId (required), sessionId, preparation, agentAddress, +1 more.

prepare_perp_orderA
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_capacity_signature_gated; API: /api/v1/agent/futures/prepare-order; authority: /api/v1/agent/status] Prepare an exact zero-custody Hyperliquid order for local EIP-712 signing. FarmDash returns the action, typed data, nonce, expiry and intent hash but never signs or submits. New-risk preparation requires a fresh execution-ready analysis; reduce-only preparation remains authority-bound at execution.

Inputs: coin (required), size (required), isBuy (required), price (required), +13 more.

prepare_virtuals_tender_v2A
Idempotent

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Prepare a non-spendable ACP V2 evidence draft. FarmDash revalidates simulation ownership and derives output/gas economics; the caller sends only a local task commitment and disclosure manifest. This tool cannot authorize, create, fund, or settle an ACP job.

Inputs: sessionId (required), selectionId (required), agentAddress (required), sessionToken (required), +8 more.

query_trail_heatA
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/trail-heat-query; authority: /api/v1/agent/status] DISCOVERY HEURISTIC query (NOT canonical Trail Heat): filter/sort/paginate the catalog editorial heuristic scores. Scores share the directory and export discovery_snapshot contract (catalog-editorial-v2), with campaign evidence and unknown source freshness. For quantitative DeFiLlama-backed scores use get_trail_heat.

Inputs: hot, chain, limit, offset, +6 more.

recommend_delta_hedgeB
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/hedge; authority: /api/v1/agent/status] Recommend short-perp hedge notionals for spot farming exposure. Produces a handoff only; it does not execute.

Inputs: spotExposure (required), riskPreference, marketConditions, volatilityRegime.

record_virtuals_tender_funding_v2A
Destructive

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Durably record canonical Base transaction hashes after the customer wallet submits an existing ACP V2 funding reservation. This action cannot submit a transaction itself.

Inputs: tenderId (required), sessionId (required), agentAddress (required), sessionToken (required), +2 more.

register_execution_delegationC

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Register a user-signed bounded delegation grant. The agent runtime gains bounded autonomous execution authority within signed policy limits.

Inputs: signature (required), policy, delegation, ownerAddress, +2 more.

release_autonomous_reservationA
Idempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Release an active autonomous budget reservation if signing or broadcast is aborted before on-chain submission.

Inputs: sessionId (required), ownerAddress (required), reservationId (required), reason.

request_approval_payloadA
Read-onlyIdempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/approval-payload; authority: /api/v1/agent/status] Approve: build the EIP-712 IntentApproval payload that the human approver signs before submit_signed_approval.

Inputs: intentId (required), approverAddress (required), decision, sessionId, +3 more.

request_execution_delegationB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Request an exact EIP-712 / ERC-8196 bounded delegation payload for user wallet approval. The user signs one scoped delegation so their agent runtime can execute within those bounds.

Inputs: policy, chainId, maxGasUsd, validAfter, +15 more.

request_human_approvalB
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/approve; authority: /api/v1/agent/status] Approve: submit a signed EIP-712 human approval or rejection for an intent. This compatibility tool is signed-only for normal callers.

Inputs: nonce (required), intentId (required), expiresAt (required), signature (required), +10 more.

reserve_virtuals_tender_funding_v2B

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Atomically reserve one role budget against the exact customer-signed ACP V2 cap after FarmDash verifies the on-chain job and proposed budget. This accounting action cannot approve tokens or spend the customer wallet.

Inputs: role (required), tenderId (required), sessionId (required), amountUnits (required), +3 more.

resolve_defi_intentC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/intent; authority: /api/v1/agent/status] Compatibility: resolve a legacy typed DeFi intent. New state-changing actions should create a durable FarmDashIntent and use the policy/simulation/approval lifecycle.

Inputs: action (required), chainId (required), tokenIn (required), amountIn (required), +9 more.

resume_autopilotB

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Resume the backend cycle scheduler for a paused agent session. Recommendations remain separately gated from transaction execution.

Inputs: sessionId (required), agentAddress (required), sessionToken (required).

revoke_execution_authorityC
DestructiveIdempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/authority; authority: /api/v1/agent/status] Revoke active bounded execution authority for an owner or agent session.

Inputs: grantId (required), reason, sessionId, ownerAddress, +3 more.

revoke_session_keyC
Destructive

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Revoke any active session key grants for this agent session.

Inputs: sessionId (required), agentAddress (required), sessionToken (required), ownerAddress.

run_risk_sentinelC
Read-onlyIdempotent

[Capability: paid_ready; tier: pioneer_plus; API: /api/v1/agent/risk-sentinel; authority: /api/v1/agent/status] Run FarmDash Risk Sentinel on a route or manual health/net-edge inputs before execution.

Inputs: gasUsd, toToken, protocol, slippage, +14 more.

run_workflowB

[Capability: available_with_tier_gate; tier: scout_catalog_pioneer_runs; API: /api/v1/agent/orchestrator; authority: /api/v1/agent/status] Create a guarded workflow run record for a session. Execution still requires each owning skill and explicit user confirmation.

Inputs: sessionId (required), workflowId (required), agentAddress (required), sessionToken (required), +1 more.

scan_funding_ratesB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_preview_pioneer_full; API: /api/v1/agent/futures/scan-funding; authority: /api/v1/agent/status] Scan Hyperliquid perp funding rates across the venue. Takes no parameters.

Inputs: no named parameters.

scan_market_conditionsB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: scout_plus; API: /api/v1/agent/futures/market-conditions; authority: /api/v1/agent/status] Technical snapshot (EMA, RSI, MACD, etc.) for one perp asset.

Inputs: coin (required).

select_execution_continuationA
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/continuation; authority: /api/v1/agent/status] Select a server-issued execution continuation as read-only context. Returns canonical next_operation parameters and missing prerequisites; it makes no quote, preparation, admission, debit, signing, or submission call and grants no execution authority. Unknown or expired IDs fail closed.

Inputs: continuationId (required).

select_virtuals_provider_plan_v2B
Read-onlyIdempotent

[Capability: available_when_deployment_prerequisites_are_met; tier: x402_or_subscription_with_authenticated_session; API: /api/v1/agent/virtuals-acp; authority: /api/v1/agent/status] Select a deterministic live-registry ACP V2 committee with one eligible provider per fixed role and three distinct verified controller identities. This selection-only action cannot authorize or spend funds.

Inputs: sessionId (required), agentAddress (required), sessionToken (required), accountAddress.

session_heartbeatB
Idempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/session; authority: /api/v1/agent/status] Keep an agent session alive. Extends expiry by 24h.

Inputs: sessionId (required), agentAddress (required), sessionToken (required).

session_key_statusC
Read-onlyIdempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/autopilot; authority: /api/v1/agent/status] Check the active session key grant and autonomous execution status.

Inputs: sessionId (required), agentAddress (required), sessionToken (required), ownerAddress.

simulate_intentB
Idempotent

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/simulate; authority: /api/v1/agent/status] Readiness probe for the generic intent simulation gate. Public callers cannot record trusted simulation evidence; until a server verifier is configured this returns simulation_verifier_unconfigured. Tool discovery does not grant simulation authority.

Inputs: intentId (required), sessionId, simulation, agentAddress, +1 more.

simulate_mee_intentB
Idempotent

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/mee; authority: /api/v1/agent/status] Preview declared-input cross-chain economics for a MEE-shaped intent. The result is non-authoritative and cannot be submitted while the coordinator gate is disabled.

Inputs: id (required).

simulate_pointsC
Read-onlyIdempotent

[Capability: available; tier: pioneer_plus; mock_supported: false; API: /api/v1/agent/simulate-points; authority: /api/v1/agent/status] Project FarmScore for a hypothetical configuration. Returns estimated points, gas, and speculative USD value.

Inputs: txCount (required), daysActive (required), protocolPoints (required), totalVolumeUsd (required), +3 more.

simulate_swap_executionA
Idempotent

[Capability: available_with_tier_gate; tier: scout_quote_signature_and_simulation_gated; API: /api/v1/simulate; authority: /api/v1/agent/status] Compatibility: run the legacy mandatory swap simulation. Lifecycle-native agents should use simulate_intent on a FarmDashIntent instead.

Inputs: intentId (required), walletAddress (required).

simulate_transaction_riskC
Read-onlyIdempotent

[Capability: available; tier: scout_limited_pioneer_full; API: /api/v1/agent/camp-guard; authority: /api/v1/agent/status] Apply Camp Guard transaction policy checks to an unsigned transaction. This is not an RPC simulation.

Inputs: transaction (required), walletAddress, expectedTransaction.

simulate_yield_strategyA
Read-onlyIdempotent

[Capability: available; tier: scout_preview_pioneer_full; API: /api/v1/agent/simulate-yield; authority: /api/v1/agent/status] Read-only independent yield alternatives with exact pool, chain and asset identities. Missing rates/costs remain null; hypothetical yields and rewards never enter observed totals. Returns WAIT analysis-only until wallet, policy, route, freshness and all-in economics are verified.

Inputs: capitalUsd (required), costs, protocols, stayPutApy, +4 more.

submit_mee_intentC
Destructive

[Capability: preparation_only; tier: syndicate; API: /api/v1/agent/mee; authority: /api/v1/agent/status] Capability probe for MEE submission. Currently returns typed feature_not_ready until a production coordinator and authoritative cross-chain receipt verifier are deployed.

Inputs: id (required), nonce (required), signature (required).

submit_signed_approvalB

[Capability: preparation_only; tier: session_bound; API: /api/v1/agent/intents/{intentId}/approve; authority: /api/v1/agent/status] Approve: submit a signed EIP-712 IntentApproval payload produced by request_approval_payload. This does not prepare or execute the intent.

Inputs: nonce (required), intentId (required), expiresAt (required), signature (required), +10 more.

verify_delegationB
Read-onlyIdempotent

[Capability: available_with_tier_gate; tier: pioneer_plus; API: /api/v1/agent/delegation; authority: /api/v1/agent/status] Verify Hyperliquid API wallet delegation for guarded venue-order workflows. Verification does not bypass live status, research, risk, or signature gates.

Inputs: agentAddress (required), ownerAddress (required), apiWalletAddress (required).

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    C2.6/5.0

    Scored across 94 tools

    Disambiguation2/5

    The set has many overlapping tools across intent, swap, simulation, approval, and research lanes, including compatibility aliases and near-duplicate probes. An agent must parse long descriptions to distinguish tools like get_agent_performance vs get_position_health, get_trail_heat vs query_trail_heat, and request_human_approval vs submit_signed_approval.

    Naming Consistency3/5

    Most tools use snake_case and a verb_noun pattern, but there are notable deviations such as agent_onboard, ledger_realized_pnl, session_key_status, and inconsistent Trail Heat spellings. Version suffixes like _v2 and legacy/compatibility naming also reduce predictability.

    Tool Count1/5

    94 tools is far beyond a practical agent-facing tool set and includes many compatibility, preparation-only, feature-gated, or probe tools. Even for a broad DeFi agent OS, this count creates severe selection overhead and suggests many tools should be consolidated or removed.

    Completeness3/5

    The surface covers a wide range of domains—research, intents, swaps, perps, risk, sessions, autopilot, workflows, and Virtuals ACP—but several lifecycle paths are intentionally non-functional or preparation-only and therefore dead ends. There are also notable gaps such as perp order reconciliation referenced by cancel_perp_order but not exposed as its own tool, and limited generic intent inspection/cancellation.

    Maintenance

    ActivityActive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      C
      quality
      D
      maintenance
      Self-hosted wallet MCP server for AI agents. Provides 42 tools for multi-chain crypto operations: transfers, token management, DeFi (swap, lend, stake, bridge, perp), NFT, smart contracts, and x402 payments. Supports EVM and Solana with policy engine, spending limits, and human approval.
      60
      29
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Agent-native, self-hosted MCP server for crypto trading and DeFi management. Enables agents to query balances, execute trades, and manage positions with a policy engine and secure key storage.
      8
      Apache 2.0
    • A
      license
      C
      quality
      C
      maintenance
      Web3 MCP proxy server for AI agents: EVM execution, DeFi swaps, bridges, advanced orders, market data, wallet management, and confirmation-gated writes.
      100
      85 npm
      1
      MIT