Hive Gateway
Unified gateway hosting 5 Hive Civilization MCP servers (evaluator, trade, depin, compute-grid…
我該用這個嗎
品質與安全性
發現項目(25)
- LOW在 dispute__dispute_route 中
- LOW在 evaluator__evaluator_submit_job 中
- LOW在 evaluator__evaluator_attest_job 中
- LOW在 compute-grid__computegrid_list_agents 中
- LOW在 compute-grid__computegrid_list_agents 中
- LOW在 compute-grid__computegrid_get_capacity 中
- LOW在 compute-grid__computegrid_get_capacity 中
- LOW在 compute-grid__computegrid_verify_proof 中
- LOW在 compute-grid__computegrid_verify_proof 中
- LOW在 morph__morph_brood_pending_approvals 中
根據工具定義與協定合規性的自動化分析。
上下文成本
這是每次將伺服器的工具載入模型上下文時所消耗的約略 token 數量。數量越高,可用於其他工作的注意力就越少。
安裝
一鍵安裝
將以下內容加入你的 `claude_desktop_config.json` 檔案:
{
"mcpServers": {
"hive-mcp-gateway": {
"url": "https://hive-mcp-gateway.onrender.com/mcp"
}
}
}遠端端點
https://hive-mcp-gateway.onrender.com/mcpstreamable-httphttps://hive-mcp-gateway.onrender.com/gateway/mcpstreamable-http它能做什麼
工具清單
工具(38)
🟢evaluator__evaluator_get_fees
Get the live evaluator fee schedule (3 tiers, settlement currencies, recipient addresses, ERC-8183 / Virtuals ACP v2.0 spec). No auth required.
輸入結構描述
{
"type": "object",
"properties": {}
}🟡evaluator__evaluator_submit_job(tier, job_value_usdc, subject_did, submitter_did, context)
Submit a job for evaluation. Choose tier (simple, evaluation, arbitration). Job value is quoted in USDC; fee = max($0.05, value * tier_bps / 10000). Returns job_id and quoted fee.
輸入結構描述
{
"type": "object",
"properties": {
"tier": {
"type": "string",
"description": "'simple' (0.5%), 'evaluation' (1.0%), or 'arbitration' (2.0%)"
},
"job_value_usdc": {
"type": "number",
"description": "Notional job value in USDC"
},
"subject_did": {
"type": "string",
"description": "DID of the agent or output being evaluated"
},
"submitter_did": {
"type": "string",
"description": "DID of the submitting agent"
},
"context": {
"type": "string",
"description": "Free-form context for the evaluator (max 4 KB)"
}
},
"required": [
"tier",
"job_value_usdc",
"subject_did",
"submitter_did"
]
}🟢evaluator__evaluator_get_job(job_id)
Retrieve evaluation status, verdict, and attestation for a previously-submitted job.
輸入結構描述
{
"type": "object",
"properties": {
"job_id": {
"type": "string",
"description": "Job ID returned from evaluator_submit_job"
}
},
"required": [
"job_id"
]
}⚪evaluator__evaluator_attest_job(job_id, signature)
Trigger settlement and emit the on-chain attestation for a completed job. Settles to the Hive Safe Treasury on the chain selected at submission. Requires EIP-3009 signature for Base/Ethereum.
輸入結構描述
{
"type": "object",
"properties": {
"job_id": {
"type": "string",
"description": "Job ID returned from evaluator_submit_job"
},
"signature": {
"type": "string",
"description": "EIP-3009 signature (required for EVM chains)"
}
},
"required": [
"job_id"
]
}🟢trade__trade_get_fees
Get the cross-border invoice fee schedule and SWIFT wire comparison (Hive vs. typical SMB wire fees, FX spreads, settlement times).
輸入結構描述
{
"type": "object",
"properties": {}
}🟢trade__trade_get_listing
Get the public listing metadata (target user, fee schedule, settlement currencies/chains, cumulative volume, council origin).
輸入結構描述
{
"type": "object",
"properties": {}
}🟡trade__trade_create_invoice(buyer_did, supplier_did, amount_usd, settle_currency, settle_chain, ...)
Create a cross-border invoice. Buyer (SMB) and supplier exchange DIDs; invoice amount + currency + chain selected at creation. Fee auto-computed from tier.
輸入結構描述
{
"type": "object",
"properties": {
"buyer_did": {
"type": "string",
"description": "DID of the SMB buyer"
},
"supplier_did": {
"type": "string",
"description": "DID of the overseas supplier"
},
"amount_usd": {
"type": "number",
"description": "Invoice amount in USD (max $250,000 for MVP)"
},
"settle_currency": {
"type": "string",
"description": "USDT or USDC"
},
"settle_chain": {
"type": "string",
"description": "base, ethereum, or solana"
},
"supplier_payout_address": {
"type": "string",
"description": "Supplier's receiving address on the chosen chain"
},
"memo": {
"type": "string",
"description": "Free-form invoice memo (PO number, goods description)"
}
},
"required": [
"buyer_did",
"supplier_did",
"amount_usd",
"settle_currency",
"settle_chain",
"supplier_payout_address"
]
}🟢trade__trade_get_invoice(invoice_id)
Retrieve invoice status, settlement transaction hash, and dispute history.
輸入結構描述
{
"type": "object",
"properties": {
"invoice_id": {
"type": "string",
"description": "Invoice ID returned from trade_create_invoice"
}
},
"required": [
"invoice_id"
]
}⚪trade__trade_dispute_invoice(invoice_id, claimant_did, reason, evidence_url)
Open a dispute on an invoice. Routes to HiveLaw arbitration if buyer and supplier cannot resolve. Settlement is held in escrow until resolution.
輸入結構描述
{
"type": "object",
"properties": {
"invoice_id": {
"type": "string",
"description": "Invoice ID"
},
"claimant_did": {
"type": "string",
"description": "DID of the disputing party (buyer or supplier)"
},
"reason": {
"type": "string",
"description": "Dispute reason"
},
"evidence_url": {
"type": "string",
"description": "Optional URL to supporting documents"
}
},
"required": [
"invoice_id",
"claimant_did",
"reason"
]
}🟢depin__depin_list_providers(provider_category, region)
List DePIN provider listings. Filter by category (storage, compute, gpu, bandwidth, energy, sensor, wireless), region, or capacity. No auth required.
輸入結構描述
{
"type": "object",
"properties": {
"provider_category": {
"type": "string",
"description": "storage | compute | gpu | bandwidth | energy | sensor | wireless"
},
"region": {
"type": "string",
"description": "Filter by region (e.g. us-east, eu-west)"
}
}
}🟡depin__depin_create_listing(agent_id, provider_category, unit_rate_usdc, unit_label, operator_did, ...)
List physical infrastructure capacity (storage TB, compute cores, GPU VRAM, bandwidth Mbps, sensor sample rate, etc.). 22 metadata fields supported. Match fee 0.15%.
輸入結構描述
{
"type": "object",
"properties": {
"agent_id": {
"type": "string",
"description": "Operator agent ID"
},
"provider_category": {
"type": "string",
"description": "storage | compute | gpu | bandwidth | energy | sensor | wireless"
},
"unit_rate_usdc": {
"type": "number",
"description": "Price per unit in USDC"
},
"unit_label": {
"type": "string",
"description": "Pricing unit, e.g. 'per TB-month', 'per GPU-hour'"
},
"operator_did": {
"type": "string",
"description": "Operator DID for trust scoring"
},
"payout_address": {
"type": "string",
"description": "Settlement address"
},
"region": {
"type": "string",
"description": "Geographic region"
},
"capacity_gb": {
"type": "number",
"description": "Capacity for storage providers (GB)"
},
"throughput_mbps": {
"type": "number",
"description": "Throughput for bandwidth providers (Mbps)"
},
"gpu_model": {
"type": "string",
"description": "GPU model for GPU providers"
},
"vram_gb": {
"type": "number",
"description": "VRAM for GPU providers (GB)"
},
"kind": {
"type": "string",
"description": "Listing kind (default: depin_provider)"
}
},
"required": [
"agent_id",
"provider_category",
"unit_rate_usdc",
"unit_label",
"operator_did",
"payout_address"
]
}🟢depin__depin_get_match_fee
Get the current DePIN marketplace match fee (currently 0.15%). Returned alongside settlement currencies and chains.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢compute-grid__computegrid_list_agents
List the 15-agent compute grid fleet across all 6 driver types. Returns agent type, count, and revenue model. No auth required.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢compute-grid__computegrid_get_capacity
Read-only capacity view from the Capacity Listener fleet. Per spec section 8: NO bids, NO hedges, NO positions, NO derivatives — pure read-only telemetry.
輸入結構描述
{
"type": "object",
"properties": {}
}🟡compute-grid__computegrid_verify_proof(job_id, driver, claimed_output_hash, submitter_did)
Submit a compute job for verification by the Verification Fleet (4 agents). Returns Groth16-style proof. $0.001/proof in USDC.
輸入結構描述
{
"type": "object",
"properties": {
"job_id": {
"type": "string",
"description": "Job ID to verify"
},
"driver": {
"type": "string",
"description": "Source driver: ionet | render | akash | aleo | custom"
},
"claimed_output_hash": {
"type": "string",
"description": "SHA-256 of claimed output"
},
"submitter_did": {
"type": "string",
"description": "DID of the submitting agent"
}
},
"required": [
"job_id",
"driver",
"claimed_output_hash",
"submitter_did"
]
}🟢morph__morph_audit_recent(limit)
Read recent rows from the polymorphic audit log: outcome, shape, counterparty hash, token id, revenue. Anonymized counterparty (hashed). No auth.
輸入結構描述
{
"type": "object",
"properties": {
"limit": {
"type": "number",
"description": "Max rows to return (default 50, max 500)"
}
}
}🟢morph__morph_get_identity(mii_id)
Look up a Morph Identity Index (MII) record. Returns the polymorphic envelope: capabilities, current shape, supermodel parent, trust score.
輸入結構描述
{
"type": "object",
"properties": {
"mii_id": {
"type": "string",
"description": "MII identifier"
}
},
"required": [
"mii_id"
]
}🟢morph__morph_list_supermodels
List all supermodels (W1 through W19): id, name, role, lane, lead shape, tagline. The supermodel directory is the canonical lineage for every spawned brood variant.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢morph__morph_get_supermodel(name_or_id)
Fetch a single supermodel by name or id (e.g. 'MONROE', 'W1'). Returns full role, lane, lead-shape, tagline, address, and brood conversion summary.
輸入結構描述
{
"type": "object",
"properties": {
"name_or_id": {
"type": "string",
"description": "Supermodel name (e.g. MONROE) or id (e.g. W1)"
}
},
"required": [
"name_or_id"
]
}🟢morph__morph_carousel
Read the polymorphic carousel: primary shape + the 7 verticals (Merchant / Provenancer / Attestor / Refunder / Creditor / Oracle / Guardian) and contrails for each.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢morph__morph_brood_pending_approvals
List brood variants currently awaiting human/operator approval before promotion. Read-only — approval itself is performed via the private hivemorph operator surface.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪morph__morph_brood_conversion
Per-variant conversion table: parent supermodel, variant id, kit version, offers shown, settles, revenue (USDC), first-offer timestamp.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢morph__morph_brood_all
List all brood variants across every supermodel. Useful for fleet-wide population scans.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪morph__morph_brood_conversion_leaderboard
Conversion leaderboard across all brood variants ranked by settle rate × revenue. Drives the auto-cull and auto-promote signals.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢morph__morph_brood_for_supermodel(supermodel)
List brood variants for a specific supermodel (e.g. MONROE / W1).
輸入結構描述
{
"type": "object",
"properties": {
"supermodel": {
"type": "string",
"description": "Supermodel name or id"
}
},
"required": [
"supermodel"
]
}🟢morph__morph_spawn_monitor_scan
Run the spawn-monitor scan: detect supermodels eligible for new brood spawns based on conversion gaps, opportunity surface, and saturation. READ-ONLY — does not actually spawn.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪morph__morph_money_flavor_probe
Probe the money-flavor classifier on the most recent settlement window: asset class histogram, chain class histogram, flow class histogram, arb-high count, refused count, non-USDC share.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪morph__morph_money_flavor_stats
Rolling aggregate money-flavor statistics (default 1 hour window): asset class, chain class, flow class distributions, plus arb/refused/non-USDC ratios. Drives the CLEAN-MONEY gate.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢morph__morph_auto_cull_scan
Run the auto-cull dry-run scan: which brood variants the system would cull based on conversion floors, ROI thresholds, and saturation. READ-ONLY — does not actually cull.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢insurance-broker__insurance_products
List all available coverage products across providers (Nexus Mutual, Sherlock, Risk Harbor, InsurAce). Returns provider, type, capacity, and current cost-of-coverage where the upstream exposes it. Real third-party listings — Hive is broker-only and does not underwrite.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪insurance-broker__insurance_quote(protocol, cover_amount_usd, duration_days, provider)
Route a quote request to one or all underwriters. Hive forwards the request to the underwriter's own quote endpoint and returns the response verbatim. Hive does NOT bind coverage, accept premium, or take custody.
輸入結構描述
{
"type": "object",
"properties": {
"protocol": {
"type": "string",
"description": "Protocol/product identifier (e.g. '2' for Nexus Mutual Aave v2, or the productId from /products)"
},
"cover_amount_usd": {
"type": "number",
"description": "Notional coverage in USD"
},
"duration_days": {
"type": "number",
"description": "Coverage duration in days (1–365)"
},
"provider": {
"type": "string",
"description": "Provider key. If omitted, quote routes to all four providers. One of: nexus_mutual, sherlock, risk_harbor, insurace"
}
},
"required": [
"protocol",
"cover_amount_usd",
"duration_days"
]
}🟢insurance-broker__insurance_today
24-hour rollup: total listing count + top providers by capacity. Returns request count and quote count for the rolling window.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢dispute__dispute_check(address, chain)
Given a counterparty address + chain, return public dispute history, active arbitration cases, and on-chain reversal-pattern flags. Sources: Kleros / UMA / Reality.eth subgraphs (with direct on-chain RPC fallback) plus Etherscan-family transaction scans. Observational only — Hive does NOT judge, freeze, or enforce.
輸入結構描述
{
"type": "object",
"properties": {
"address": {
"type": "string",
"description": "0x-prefixed counterparty address (20 bytes)"
},
"chain": {
"type": "string",
"description": "ethereum | base | arbitrum"
}
},
"required": [
"address",
"chain"
]
}🟢dispute__dispute_providers
List available third-party arbitration protocols (Kleros, UMA Optimistic Oracle, Reality.eth) with current case load, intake URLs, and jurisdiction model. Hive is not one of them — Hive only surfaces signal.
輸入結構描述
{
"type": "object",
"properties": {}
}⚪dispute__dispute_route(description, amount_usd, jurisdiction)
Given a case description, disputed amount in USD, and optional jurisdiction preference, return ranked arbitration provider options. NO automatic filing — Hive returns options + intake URLs only. The disputing party files directly with the chosen provider.
輸入結構描述
{
"type": "object",
"properties": {
"description": {
"type": "string",
"description": "Free-form case description (max 2000 chars)"
},
"amount_usd": {
"type": "number",
"description": "Disputed amount in USD"
},
"jurisdiction": {
"type": "string",
"description": "Optional preference (e.g. 'decentralized', 'optimistic', 'crowdsourced')"
}
},
"required": [
"description",
"amount_usd"
]
}🟢dispute__dispute_today
24-hour rollup: flagged-counterparty count + top arbitration providers by active case load.
輸入結構描述
{
"type": "object",
"properties": {}
}🟢audit-readiness__audit_readiness_score(organization_country, jurisdictions, data_volume_records, agent_count, monthly_inference_calls, ...)
Compute a multi-jurisdictional AI compliance readiness score for an organization. Returns penalty exposure (EUR + USD), specific compliance gaps citing the regulation article, recommended audit tier (STARTER/STANDARD/ENTERPRISE/FEDERAL), and the nearest enforcement deadline. Penalty math sources EU AI Act Art 99, Colorado AI Act SB 24-205, CCPA, Cal SB 942, NYC LL 144, HIPAA. Free, no auth, rate-limited 10/IP/hr.
輸入結構描述
{
"type": "object",
"properties": {
"organization_country": {
"type": "string",
"description": "ISO 3166-1 alpha-2 country code (e.g. \"US\", \"DE\", \"FR\", \"GB\")."
},
"jurisdictions": {
"type": "array",
"items": {
"type": "string"
},
"description": "Where the system operates: [\"EU\", \"US-CO\", \"US-CA\", \"US-NY\", ...]. Drives which regulations apply."
},
"data_volume_records": {
"type": "integer",
"description": "Total records processed (drives CCPA / GDPR scoping)."
},
"agent_count": {
"type": "integer",
"description": "Number of distinct AI agents in production."
},
"monthly_inference_calls": {
"type": "integer",
"description": "Inference call volume per month (drives tier selection)."
},
"sectors": {
"type": "array",
"items": {
"type": "string"
},
"description": "Industries: [\"finance\", \"healthcare\", \"employment\", \"education\", \"lending\", \"insurance\", \"criminal_justice\", \"biometric\", \"critical_infrastructure\"]."
},
"frameworks": {
"type": "array",
"items": {
"type": "string"
},
"description": "Regulations to score: [\"eu_ai_act\", \"co_ai_act\", \"ccpa\", \"ca_sb942\", \"nyc_ll144\", \"hipaa\", \"gdpr\", \"nist_ai_rmf\"]."
},
"company": {
"type": "string",
"description": "Organization name (optional; populates the assessment record)."
}
},
"required": [
"organization_country",
"jurisdictions",
"data_volume_records",
"agent_count",
"monthly_inference_calls",
"frameworks"
]
}🟢audit-readiness__audit_get_tier_pricing
Returns the four published HiveAudit tier prices and bracket thresholds: STARTER ($500, <$500K exposure), STANDARD ($1,500, <$5M), ENTERPRISE ($2,500, <$50M), FEDERAL ($7,500/yr, ≥$50M or federal agency). Inlined, no backend call.
輸入結構描述
{
"type": "object",
"properties": {}
}社群
證據