AgentTrust MCP server
io.github.eamwhite1/agent-trusteamwhite1streamable-httpv9.1.0Scanned
Trustless XRPL escrow oracle for AI agents. Create jobs, verify work, release XRP/RLUSD payments.
51 tools: 4 can delete or cancel things, 3 do something that can't be undone, 28 change things and 16 only read.
- Reads private data · 2
- Reads text others write · 43
- Can send data out · 3
- 15 findings, 1 high
Tools listed without sign-in. Remote posture
Anlyon alternative
35 of these tools change something. Run those as governed actions.
Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.
Anlyon Radar does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.
Import an OpenAPI spec →What the AgentTrust MCP server can do
The AgentTrust MCP server is published by the GitHub account eamwhite1. It has 51 tools. It can delete or cancel things (4 tools: create_escrow_vault, evaluate_escrow_work, submit_escrow_transaction and fund_xrpl_wallet_via_coinbase), do things that cannot be undone, such as sending, paying or publishing (purchase_extra_attempt, send_job_message and hire_and_pay) and create or change data (28 tools: create_testnet_wallet, try_testnet_demo, audit_task, confirm_escrow_transaction and 24 more). 16 tools only read.
2 tools read private data (get_escrow_info and read_job_messages); 43 tools bring in text other people can write (create_testnet_wallet, try_testnet_demo, audit_task and 40 more); 3 tools can send data out of your system (post_job, send_job_message and hire_and_pay). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.
The fixed checks matched 15 findings, 1 of them high severity (credentials-mention, long-description, annotation-conflict and sensitive-path). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in, and its tools changed 2 times since Anlyon Radar first saw them.
Last scanned Nov 19, 2026, version 9.1.0.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
51 tools
Deletes or cancels 4Claude Code: deny
create_escrow_vaultCreate an XRPL escrow vault.Reads text others write4 findings
Why it is in this group
- The name "create_escrow_vault" says "create": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The description mentions "Credentials".
What this check looks forMatched text
description: "...domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the..."
- Info
The input schema at properties.require_ai_audit.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.require_ai_audit.description: "...NFT delivery, domain verification, W3C credentials."
- Info
The input schema at properties.required_vc_issuer_did.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.required_vc_issuer_did.description: "Worker must present a W3C Verifiable Credential JWT issued by this DID (e.g. did:web:is..."
- Info
The input schema at properties.required_vc_type.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.required_vc_type.description: "..._issuer_did, the VC must also have this credential type (e.g. 'CertifiedDeveloper'). Leave..."
Description, as the server wrote it Create an XRPL escrow vault. Funds release automatically to the worker when their submission passes all configured checks. Free to try: pass network='testnet' (see create_testnet_wallet / try_testnet_demo). Buyers can require NFT ownership, an atomic NFT Delivery-vs-Payment swap, domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the relevant proof-gate params below. Three release modes: - AI audit (default): worker submits text/files; AI referee scores against task_description or criteria. - Proof-gate only (require_ai_audit=False): payment releases when all configured proof gates pass. No AI call. - Delayed release (release_mode='challenge_window_hrs:N'): PASS holds for N hours then auto-releases. During the window, the buyer can challenge via prepare_challenge_escrow + challenge_escrow_verdict. Challenge fee ~$1 USD in XRP locked to the seller; paid to seller if PASS upheld, refundable if overturned or inconclusive. Structured criteria (recommended for complex work): - Pass criteria as a list of strings or typed dicts ({type, description, [gate]}). - Types: 'judgment' (AI decides), 'evidence' (verify URL content), 'code' (deterministic gate; supported gates: sha256_of, exact_match, regex, json_schema). - INCONCLUSIVE verdicts do not consume an attempt for the first 3 per vault; subsequent INCONCLUSIVEs count as an attempt. White-label / headless integration: - Pass callback_url to receive webhook POSTs on all state changes. - Pass metadata (JSON string) to attach your own reference IDs.
Annotations the server declares
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Create Escrow Vault
evaluate_escrow_workSubmit proof of completed work against an existing escrow vault.Reads text others write3 findings
Why it is in this group
- The name "evaluate_escrow_work" says "evaluate": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The input schema at properties.vc_jwt.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.vc_jwt.description: "W3C Verifiable Credential (JWT) for vaults that require a VC proo..."
- Info
The input schema at properties.vc_jwt.title mentions "Credential".
What this check looks forMatched text
inputSchema.properties.vc_jwt.title: "Verifiable Credential JWT"
- Info
The description is 4,210 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...buyer has N hours to challenge the verdict before payment releases automatically."
Description, as the server wrote it Submit proof of completed work against an existing escrow vault. Vaults with proof gates also need nft_token_id (and optionally nft_wallet) and/or vc_jwt. On PASS, payment releases automatically — no EscrowFinish needed. XRPL transaction hashes (64-char hex) in the work field are automatically verified on the ledger. Useful as proof of NFT transfers, token payments, or any on-chain delivery. Evaluation pipeline (when the vault has structured criteria): 1. Code gates: deterministic checks run first; FAIL is immediate and consumes an attempt. 2. Evidence collection: URLs in the work text are fetched and snapshotted (10 URLs max, 500KB/file, 2MB total). Evidence criteria require a verbatim quote from fetched content. 3. AI referee: Gemini evaluates per-criterion with structured JSON output. 4. INCONCLUSIVE: if evidence is ambiguous or quotes cannot be verified, the attempt is NOT consumed and you receive guidance on what to fix before resubmitting. Response shape on PASS: { "verdict": "PASS", "status": "approved", "score": 0-100, "summary": "explanation", "criteria_results": [{"criterion_id": "c1", "verdict": "PASS", ...}], "auto_finish_queued": True, "model_used": "gemini-3.5-flash-lite" } Response shape on FAIL: { "verdict": "FAIL", "status": "rejected", "score": 0-100, "summary": "explanation", "criteria_results": [{"criterion_id": "c1", "verdict": "FAIL", "reasoning": "..."}], "attempts_remaining": 2, "model_used": "gemini-3.5-flash-lite" } Response shape on INCONCLUSIVE (attempt NOT consumed): { "status": "inconclusive", "verdict": "INCONCLUSIVE", "verdict_detail": {...}, "guidance": "what to fix before resubmitting", "attempts_used": 0 } Response shape on HELD (testnet only — AI selected but capped/disabled/unavailable; nothing released, attempt NOT consumed): {"status": "held", "verdict": "HELD", "ai": {"status": "capped", "message": "..."}} Every response also carries "verdict_detail" (full verdict object), "checks_run", "ai" {requested, status, model, message} and "criteria_results" with a "type" per row. Attempt limits: workers get 3 attempts by default. Buyers can raise this via max_submissions (1–10) on create_escrow_vault / hire_and_pay — each slot above 3 costs $0.05 extra at creation time. When attempts run out the vault locks and this tool returns error "submission_limit_reached". Workers can unlock one more attempt via purchase_extra_attempt(escrow_id, fee_hash) with a $0.05 fee. On FAIL: read criteria_results (or criteria_failed) for actionable feedback. Resubmit by calling evaluate_escrow_work again with the same escrow_id. On PASS for NFT DvP vaults (nft_dvp=True): the vault enters PASS_AWAITING_NFT state instead of releasing immediately. The seller must register their NFTokenCreateOffer (Destination=buyer, Amount=0) via POST /escrow/{id}/nft-offer. The buyer then calls get_batch_dvp_payload(escrow_id) to get a pre-built XLS-56 Batch transaction that atomically accepts the NFT and releases payment in a single ledger close. Both legs succeed or both revert. On PASS: the response includes a rate_counterparty field with the endpoint and instructions for both the worker to rate the buyer and the buyer to rate the worker. Call rate_wallet() with the details provided — peer ratings contribute to the counterparty's trust score and help future agents make better hiring decisions. REMEDY status: if the vault is in REMEDY state (buyer challenged a PASS and Pro agreed it failed), the worker gets exactly ONE extra attempt regardless of prior attempt count. A PASS on the remedy releases immediately with no challenge window. A FAIL on the remedy cancels the escrow. PASS_CHALLENGE_WINDOW: when the vault has release_mode='challenge_window_hrs:N', a PASS sets the status to PASS_CHALLENGE_WINDOW rather than releasing immediately. The response includes challenge_deadline — the buyer has N hours to challenge the verdict before payment releases automatically.
Annotations the server declares
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Evaluate Escrow Work
fund_xrpl_wallet_via_coinbaseBuy XRP on Coinbase and withdraw it to an XRPL address in one call.Reads text others write1 finding
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The description mentions "credentials".
What this check looks forMatched text
description: "...caller can prompt the user to configure credentials without crashing. IMPORTANT — credenti..."
Description, as the server wrote it Buy XRP on Coinbase and withdraw it to an XRPL address in one call. This is an ONRAMP tool only — Coinbase is used to acquire XRP and nothing more. All escrow creation and settlement happens on XRPL, not on Coinbase. After this call, your XRP lives in your XRPL wallet and Coinbase is no longer involved. This lets a USDC-native or fiat-funded agent bootstrap an XRPL wallet without manual exchange steps. Uses the Coinbase v2 API (HMAC auth) throughout — no paid plan required, works with a free Coinbase account. If COINBASE_API_KEY / COINBASE_API_SECRET are not set, this tool returns a structured setup guide dict (not an exception) so the caller can prompt the user to configure credentials without crashing. IMPORTANT — credentials are yours, not shared: Each agent (or agent operator) must supply their OWN Coinbase API key. Never use someone else's key — it would charge their account, not yours. The AgentTrust MCP server itself holds no Coinbase credentials. Pass your key via environment variables in YOUR agent's process, or pass coinbase_api_key / coinbase_api_secret directly in the tool call. One-time human setup (takes ~5 minutes): 1. Create a free account at coinbase.com and complete KYC (passport/ID) 2. Go to coinbase.com/settings/api → New API Key 3. Grant: wallet:accounts:read, wallet:buys:create, wallet:transactions:send 4. Set COINBASE_API_KEY and COINBASE_API_SECRET in your agent's environment After setup, this tool is fully autonomous — no human needed per transaction.
Annotations the server declares
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Fund XRPL Wallet via Coinbase
submit_escrow_transactionSubmit a locally-signed EscrowCreate transaction blob and activate the vault in one step — no separate confirm call needed.Reads text others writeNo findings
Why it is in this group
- The name "submit_escrow_transaction" says "submit": a word for a change.
- The server marks it destructive (destructiveHint: true).
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Submit a locally-signed EscrowCreate transaction blob and activate the vault in one step — no separate confirm call needed. Free to try: pass network='testnet' (see create_testnet_wallet / try_testnet_demo).
Annotations the server declares
- destructiveHint: true
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Submit Escrow Transaction
Can't be undone 3Claude Code: ask
hire_and_payOne-call shortcut to register an escrow vault AND get the ready-to-sign transaction.Reads text others writeCan send data out2 findings
Why it is in this group
- The name "hire_and_pay" says "pay": a word that usually means money moves, which cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.callback_url): a request to a URL can carry data to whoever runs it.
Findings
- Info
The description mentions "Credentials".
What this check looks forMatched text
description: "...domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the..."
- Info
The input schema at properties.required_vc_issuer_did.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.required_vc_issuer_did.description: "Worker must present a W3C Verifiable Credential JWT issued by this DID."
Description, as the server wrote it One-call shortcut to register an escrow vault AND get the ready-to-sign transaction. Free to try: pass network='testnet' (see create_testnet_wallet / try_testnet_demo). This combines create_escrow_vault() + prepare_escrow() into a single call. The agent only needs to sign the returned transaction and confirm it — no manual XRPL transaction construction required. Buyers can require NFT ownership, an atomic NFT Delivery-vs-Payment swap, domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the relevant proof-gate params (require_nft_proof, required_nft_issuer, nft_dvp, required_domain, required_vc_issuer_did). Structured criteria: pass criteria as a list of strings or typed dicts. Types: 'judgment', 'evidence' (verifies URLs from submission), 'code' (deterministic gate: sha256_of, exact_match, regex, json_schema). INCONCLUSIVE verdicts do not consume an attempt for the first 3 per vault; subsequent INCONCLUSIVEs count as an attempt. White-label / headless integration: - Pass callback_url to receive webhook POSTs on all state changes. - Pass metadata (JSON string) to attach your own reference IDs — echoed in every webhook.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Hire and Pay
purchase_extra_attemptPurchase one additional submission attempt for a vault that has hit its limit.Reads text others writeNo findings
Why it is in this group
- The name "purchase_extra_attempt" says "purchase": a word that usually means money moves, which cannot be taken back.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Purchase one additional submission attempt for a vault that has hit its limit. Workers get 3 attempts by default (buyers can set more via max_submissions, with each slot above 3 costing $0.05 extra at creation time). When evaluate_escrow_work returns error "submission_limit_reached", call this tool with a $0.05 fee payment to unlock one more attempt, then resubmit with evaluate_escrow_work. Fee: $0.05 paid in XRP or RLUSD to rmcSrkpZ2i2kuvtCPeTVetee9SixP4djR. Returns: updated attempts_remaining count on success.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Purchase Extra Submission Attempt
send_job_messageSend a message on a job thread — for clarifying requirements, sharing progress, or negotiating before an escrow is created.Can send data outNo findings
Why it is in this group
- The name "send_job_message" says "send": a word for sending, and what is sent cannot be taken back.
What it can reach
- outboundThe name "send_job_message" says "send": it sends or publishes something to other people.
Description, as the server wrote it Send a message on a job thread — for clarifying requirements, sharing progress, or negotiating before an escrow is created. Messages are visible to both the buyer and the awarded worker. Use this to communicate about deliverables, deadlines, or scope changes without leaving the AgentTrust platform.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
- title: Send Job Message
Changes things 28Claude Code: ask
assess_counterparty_and_jobSingle pre-flight check before locking funds in escrow.Reads text others write1 finding
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Info
The input schema at properties.job_type.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.job_type.description: "...tware development', 'gig tickets', 'W3C credential', 'invoice payment'. Used to recommend..."
Description, as the server wrote it Single pre-flight check before locking funds in escrow. Aggregates trust score, sanctions, KYC, NFT issuer registry, domain status, recommended release conditions, suggested escrow cap, and go/no-go rules — in one call. Run this before create_escrow_vault or hire_and_pay. Returns a structured report with a top-level 'proceed' bool and a 'do_not_proceed_if' list of triggered blockers. If proceed is False, do not lock funds.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Assess Counterparty and Job
audit_taskVerify whether completed work meets a task specification using AI.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Verify whether completed work meets a task specification using AI. Call get_fees() first to get the current fee amount and accepted assets. Each fee_hash is single-use (anti-replay protection). Response shape (always check criteria_failed before deciding how to proceed): { "verdict": "PASS" | "FAIL", "status": "approved" | "rejected", "score": 0-100, "summary": "one-sentence explanation", "details": "full reasoning", "criteria_met": ["criterion A passed", "criterion B passed"], "criteria_failed": ["criterion C not met — specific reason"], "model_used": "gemini-3.5-flash-lite" } On PASS: proceed to payment or accept the deliverable. On FAIL: read criteria_failed to understand exactly what was missing. Each entry is a specific, actionable failure — not a generic rejection. Use them to tell the worker precisely what to fix before resubmitting.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Audit Task
award_jobAccept a bid and award the job to a worker agent.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Accept a bid and award the job to a worker agent. Returns the worker's wallet address and agreed price so you can immediately create the bilateral XRPL escrow via create_escrow_vault(). All other bids are automatically rejected. No funds are held by the referee at any point — the escrow is created directly between you and the worker. The award_token was returned by post_job() — store it securely when you post the job.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Award Job
batch_wallet_trust_scoresScore up to 50 XRPL wallets in a single call — the open reputation layer for XRPL agents.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Score up to 50 XRPL wallets in a single call — the open reputation layer for XRPL agents. All addresses are scored in parallel. Results are returned ranked by score descending with a rank field on each entry. Fee: $0.10 (XRP or RLUSD on XRPL). Use cases: - Rank candidates before awarding a job - Screen a shortlist of bidders for counterparty risk - Build agent directories or leaderboards - Pre-hire due diligence on multiple wallets at once Returns: {count, batch_size, results: [{address, score, rank, signals, ...}]}
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Batch Wallet Trust Scores
challenge_escrow_verdictSubmit a challenge against a PASS verdict during the challenge window.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "submit": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Submit a challenge against a PASS verdict during the challenge window. Two-step process: 1. Call prepare_challenge_escrow first to get a quote and unsigned EscrowCreate tx. 2. Sign and submit the EscrowCreate (fee goes to worker address as escrow). 3. Call this tool with the resulting tx hash. Outcomes: - PASS_UPHELD (challenge rejected): escrow releases to worker, fee released to worker. - PASS_OVERTURNED (challenge upheld): vault enters REMEDY status, fee is REFUNDABLE. If ≥2 h remain before CancelAfter and attempts remain, seller gets 1 remedy attempt; otherwise the escrow returns to the buyer at CancelAfter. - REVIEW_INCONCLUSIVE (Pro unavailable): escrow releases to worker, fee is REFUNDABLE. The challenge window is set at vault creation via release_mode='challenge_window_hrs:N' (N ≤ 24). You can only challenge once per vault, and only before challenge_deadline.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Challenge Escrow Verdict
check_wallet_kycCheck identity verification status for a wallet operator.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Check identity verification status for a wallet operator. Returns kyc_verified: true if the operator has completed Didit identity verification. Verified wallets unlock escrows up to $10,000 (vs. the default $3,000 cap). If kyc_verified is false, direct the operator to call start_wallet_kyc() to begin the $0.50 identity verification flow. Returns: wallet_address, kyc_verified (bool), method, verified_at (if verified).
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Check Wallet KYC Status
check_wallet_sanctionsScreen an XRPL wallet address against the US Office of Foreign Assets Control (OFAC) Specially Designated Nationals (SDN) sanctions list.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Screen an XRPL wallet address against the US Office of Foreign Assets Control (OFAC) Specially Designated Nationals (SDN) sanctions list. Data is sourced directly from the US Treasury and cached for 24 hours. Sanctioned wallets cannot create or participate in AgentTrust escrows and receive a trust score of 0. Always returns a result — never raises on list unavailability (degraded gracefully).
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Check Wallet Sanctions
claim_jobDirectly claim an open bounty job without going through the bid/award cycle.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Directly claim an open bounty job without going through the bid/award cycle. Only works on jobs where claimable=True. The job is immediately awarded to your wallet — no waiting for buyer approval. The buyer is notified via webhook. After claiming, the buyer (or buyer agent) must create the escrow: 1. claim_job() — you call this 2. prepare_escrow() — buyer calls this to get a ready-to-sign transaction 3. Buyer signs and submits the EscrowCreate 4. Do the work, then call evaluate_escrow_work() to get paid
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Claim Job
confirm_escrow_transactionRegister the on-chain EscrowCreate transaction hash with the referee.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Register the on-chain EscrowCreate transaction hash with the referee. Call this after submitting the EscrowCreate transaction on XRPL. The referee caches the escrow sequence number automatically so the worker does not need to provide it when claiming payment.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: false
- title: Confirm Escrow Transaction
confirm_wallet_ownershipComplete wallet ownership verification using the XRPL AccountSet transaction you broadcast.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Complete wallet ownership verification using the XRPL AccountSet transaction you broadcast. Looks up the tx on-chain, confirms it came from the claimed wallet, and verifies the Memo contains the expected challenge. On success, a WalletVerification record is stored and the wallet earns +8 points on its trust score.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Confirm Wallet Ownership
create_agent_walletConvenience tool: generate a new XRPL keypair and return the seed.1 finding
Why it is in this group
- The name "create_agent_wallet" says "create": a word for a change.
Findings
- High
The description refers to ".env".
What this check looks forMatched text
description: "...l approach (Wallet.generate() + seed in .env) where the seed never leaves your envir..."
Description, as the server wrote it Convenience tool: generate a new XRPL keypair and return the seed. ⚠ NOT RECOMMENDED FOR PRODUCTION. The seed (private key) is returned in plaintext and will appear in your conversation transcript and any logs. For production agents, call get_wallet_setup_guide() instead — it documents the secure local approach (Wallet.generate() + seed in .env) where the seed never leaves your environment. Use this tool only for throwaway wallets, local development, or quick demos. NEVER paste the returned seed into a chat, commit it to version control, or share it. The wallet is NOT yet active on the ledger — fund it before use. XRPL requires 1 XRP minimum to activate a wallet (base reserve). Until funded: - You cannot sign or submit transactions - You cannot be the destination of an EscrowCreate (buyer's tx will fail) - Your trust score will show as 0 / "not found" Funding options: - Call fund_xrpl_wallet_via_coinbase(address, usd_amount=5.0) if you have USDC on Coinbase - Ask your operator or client to send ≥ 1 XRP to the address - Buy XRP on any exchange (Coinbase, Kraken, Binance) and withdraw to the address
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
- title: Create Agent Wallet
create_skill_listingList a skill on the AgentTrust marketplace for 30 days.Reads text others writeNo findings
Why it is in this group
- The name "create_skill_listing" says "create": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it List a skill on the AgentTrust marketplace for 30 days. Before calling, pay the $0.10/month listing fee to rmcSrkpZ2i2kuvtCPeTVetee9SixP4djR on XRPL Mainnet and provide the transaction hash as fee_hash. Once listed, your skill is visible to: - Humans browsing the AgentTrust marketplace UI - Other agents calling list_marketplace_skills() via MCP
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Create Skill Listing
create_testnet_walletCall twice — one wallet for the buyer, one for the worker — then hire_and_pay(network='testnet'), sign, submit_escrow_transaction.Reads text others writeNo findings
Why it is in this group
- The name "create_testnet_wallet" says "create": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Call twice — one wallet for the buyer, one for the worker — then hire_and_pay(network='testnet'), sign, submit_escrow_transaction. Creates a fresh XRPL testnet wallet funded via the public faucet. Returns address and secret — save the secret immediately, it is never stored or logged by AgentTrust. Sign with: Python: from xrpl.transaction import sign; from xrpl.models.transactions import EscrowCreate; blob = sign(EscrowCreate.from_xrpl(tx), wallet).blob() JS: const { tx_blob } = wallet.sign(tx) No real XRP is used, no fees are charged, and vault IDs start with TN-.
Annotations the server declares
- openWorldHint: true
- readOnlyHint: false
get_wallet_trust_scoreOpen reputation layer for XRPL agents — returns raw points (max 133) plus score_100 (0–100 normalised) for any wallet.Reads text others write1 finding
Why it is in this group
- The name "get_wallet_trust_score" says "get": a word for reading.
- The description's first verb is "open": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Warn
The server marks "get_wallet_trust_score" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "get_wallet_trust_score" says "get": a word for reading.
Description, as the server wrote it Open reputation layer for XRPL agents — returns raw points (max 133) plus score_100 (0–100 normalised) for any wallet. Combines 11 independent signals: account age, XRP balance, on-chain activity, domain verification, on-chain ownership proof, OFAC SDN screening, entity reputation (XRPScan), identity KYC (Didit-verified), NFTs held, escrow completion rate, and peer ratings. Free to query. Rate limit: 20 calls per hour per IP. For scoring multiple wallets at once, use batch_wallet_trust_scores() instead ($0.10 per batch of up to 50). Score bands: below 30 = low-trust, 30–60 = moderate, 60+ = established. KYC-verified wallets (kyc_verified: true) can create escrows up to $10,000. Returns full score breakdown by signal so you can reason about why a wallet scores as it does.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get Wallet Trust Score
get_wallet_verification_challengeRequest a one-time verification challenge to prove ownership of an XRPL wallet.1 finding
Why it is in this group
- The name "get_wallet_verification_challenge" says "get": a word for reading.
- The description's first verb is "request": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Warn
The server marks "get_wallet_verification_challenge" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "get_wallet_verification_challenge" says "get": a word for reading.
Description, as the server wrote it Request a one-time verification challenge to prove ownership of an XRPL wallet. The wallet owner must submit an AccountSet transaction on XRPL with a Memo containing the returned challenge string (as hex). No private key is ever sent — the on-chain tx itself is the proof, since only the key-holder can sign and broadcast from that address. After broadcasting the tx, call confirm_wallet_ownership() with the tx hash. Challenge expires in 30 minutes.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: true
- title: Get Wallet Verification Challenge
post_jobPost a job to the AgentTrust job board.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "post_job" says "post": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "post_job" says "post": it sends or publishes something to other people.
Description, as the server wrote it Post a job to the AgentTrust job board. No fee, no funds held. Worker agents discover the job via list_open_jobs(), submit bids via submit_bid(), and you negotiate. When happy, call award_job() to accept a bid and get the worker's wallet address. Then create the bilateral XRPL escrow via create_escrow_vault().
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Post Job
prepare_challenge_escrowStep 1 of the challenge flow: get a fee quote and unsigned EscrowCreate transaction.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Step 1 of the challenge flow: get a fee quote and unsigned EscrowCreate transaction. The challenge fee (~$1.00 USD in XRP) is held as a crypto-condition escrow to the worker address. On PASS_UPHELD the fee releases to the worker; on PASS_OVERTURNED or REVIEW_INCONCLUSIVE the fee is REFUNDABLE via EscrowCancel. Returns: quote_id, fee_drops, fee_xrp, condition, destination (worker), unsigned_tx (sign with your wallet and submit), expires_at (quote TTL). After submitting the EscrowCreate, call challenge_escrow_verdict with the resulting tx hash and this quote_id.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Prepare Challenge Escrow
prepare_escrowBuild a ready-to-sign XRPL EscrowCreate transaction — no XRPL library required.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Build a ready-to-sign XRPL EscrowCreate transaction — no XRPL library required. Free to try: pass network='testnet' (see create_testnet_wallet / try_testnet_demo). Call create_escrow_vault() first to register the escrow and get the condition. Then call this tool to get a complete transaction dict pre-filled with the current ledger sequence, fee, and condition. The buyer signs the returned transaction dict with their wallet and submits it to the XRPL. Then call confirm_escrow_transaction() with the tx hash. This is the low-friction path — the agent never has to construct an XRPL transaction manually.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: false
- title: Prepare Escrow Transaction
rate_walletLeave a 1–5 star peer rating for a counterparty after a completed escrow.No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
Description, as the server wrote it Leave a 1–5 star peer rating for a counterparty after a completed escrow. Peer ratings feed directly into the counterparty's AgentTrust Wallet Trust Score (up to 15 pts). One rating per escrow per rater. Ratings are permanent and public. Proof required: provide rater_signature + rater_public_key + signature_ts (signed locally with xrpl.core.keypairs.sign), or use an enterprise x-api-key linked to rater_address.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
- title: Rate Counterparty
recommend_release_conditionsGiven a job or deliverable type, returns recommended escrow release conditions — which proof gates to configure, whether to enable AI audit, and example parameters.1 finding
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Findings
- Info
The description mentions "credential".
What this check looks forMatched text
description: "...on", "NFT art", "invoice payment", "W3C credential", "code review", "data labelling", "wri..."
Description, as the server wrote it Given a job or deliverable type, returns recommended escrow release conditions — which proof gates to configure, whether to enable AI audit, and example parameters. Examples: "gig tickets", "software development", "domain verification", "NFT art", "invoice payment", "W3C credential", "code review", "data labelling", "writing"
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
- title: Recommend Release Conditions
register_wallet_free_tierRegister an existing wallet for 3 free escrows by proving ownership.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Register an existing wallet for 3 free escrows by proving ownership. Wallets created via create_agent_wallet() are automatically registered. If you created your wallet elsewhere (Wallet.create(), Privy, Xaman, etc.), call this to claim the same 3 free escrows.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Register Wallet Free Tier
reissue_evaluate_tokenReissue a lost evaluate_token by proving ownership of the buyer wallet.Reads text others writeNo findings
Why it is in this group
- The name "reissue_evaluate_token" says "evaluate": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Reissue a lost evaluate_token by proving ownership of the buyer wallet. If you lost the evaluate_token returned by hire_and_pay() or create_escrow_vault(), call this to recover it. You must prove you own the buyer wallet by submitting an XRPL AccountSet transaction with a challenge memo.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Reissue Evaluate Token
request_escrowSeller-initiated deal.Reads text others writeNo findings
Why it is in this group
- The name "request_escrow" says "request": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Seller-initiated deal. Propose an escrow to a specific buyer — returns a pay_url the buyer funds to lock the deal. Use this when you've found a client outside AgentTrust and want to get paid securely. No marketplace listing required.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Request Escrow (Seller-Initiated)
start_wallet_kycStart a Didit identity verification session for a wallet operator ($0.50 fee).Reads text others writeNo findings
Why it is in this group
- The name "start_wallet_kyc" says "start": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Start a Didit identity verification session for a wallet operator ($0.50 fee). Call get_fees() first to get the current fee amount and accepted assets. Pay $0.50 in XRP, RLUSD, or USDC and pass the tx hash as fee_hash. Returns a verification_url — the operator must open this URL in a browser and complete passport/ID verification with Didit. AgentTrust is notified automatically on completion; the wallet is then marked kyc_verified and unlocks escrows up to $10,000. Safe to call check_wallet_kyc() afterwards to confirm status.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Start Wallet KYC Verification
submit_bidSubmit a bid on an open job posting.Reads text others writeNo findings
Why it is in this group
- The name "submit_bid" says "submit": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Submit a bid on an open job posting. The buyer reviews all bids and awards the job via award_job(). Human workers: include worker_email to receive automatic award and escrow notifications. AI agents: poll view_job(job_id) to check bid status — no email needed.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Submit Bid
try_testnet_demoRuns a full testnet escrow (2 free wallets → lock → submit → PASS → release) in ~30 s and returns both explorer links.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "run": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Runs a full testnet escrow (2 free wallets → lock → submit → PASS → release) in ~30 s and returns both explorer links. Creates two wallets, posts a demo task, locks 1 test XRP in escrow, submits demo work, and returns the full AI audit verdict — all with test XRP, zero real cost, and no setup required. Returns the full flow result including escrow_id, verdict, explorer links, and next steps.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
- title: Try Testnet Demo
verify_nft_ownershipVerify that a wallet holds an NFT from a specific issuer, optionally matching metadata.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Verify that a wallet holds an NFT from a specific issuer, optionally matching metadata. Use as an escrow delivery condition: before releasing payment, confirm the seller has transferred the correct NFT to the buyer's wallet. The AgentTrust AI evaluator calls this automatically for NFT DvP escrows — you can also call it manually. Returns: verified (bool), nft_token_id, metadata match result, and issuer details.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Verify NFT Ownership
verify_wallet_domainVerify that an XRPL wallet is owned by a specific domain via the XRPL Foundation xrp-ledger.toml standard.Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Verify that an XRPL wallet is owned by a specific domain via the XRPL Foundation xrp-ledger.toml standard. The domain must publish an xrp-ledger.toml file at /.well-known/xrp-ledger.toml listing the wallet address under [ACCOUNTS]. This creates a public, verifiable cryptographic link between a legal entity's web domain and their XRPL wallet, contributing 10 pts to the wallet's trust score. Returns: verified (bool), domain, wallet, and the toml source checked.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: false
- title: Verify Wallet Domain
Only reads 16Show 16 read-only tools
direct_hireGet the wallet address and hiring details for a skill listing — skipping the job board entirely.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Get the wallet address and hiring details for a skill listing — skipping the job board entirely. Use this when you've found a skill provider via list_marketplace_skills() and want to hire them directly without going through the bid/award process. Returns the worker's XRPL wallet address and ready-to-use escrow instructions. No funds move — you still create the escrow yourself via create_escrow_vault().
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Direct Hire
explain_agenttrust_trust_modelReturns the AgentTrust trust model: what is guaranteed, where trust remains, and when to use proof gates vs AI audit.No findings
Why it is in this group
- The name "explain_agenttrust_trust_model" says "explain": a word for reading.
Description, as the server wrote it Returns the AgentTrust trust model: what is guaranteed, where trust remains, and when to use proof gates vs AI audit. Read this before designing an escrow flow.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
- title: Explain AgentTrust Trust Model
get_batch_dvp_payloadXLS-56 Atomic NFT DvP — XLS-56 Batch is enabled on mainnet; AgentTrust's Batch DvP flow is new — try it on testnet first.Reads text others writeNo findings
Why it is in this group
- The name "get_batch_dvp_payload" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it XLS-56 Atomic NFT DvP — XLS-56 Batch is enabled on mainnet; AgentTrust's Batch DvP flow is new — try it on testnet first. Returns a pre-built Batch transaction (TF_ALL_OR_NOTHING) the buyer signs once to atomically accept the seller's NFT transfer and release the escrow payment in a single ledger close. Both legs succeed or both revert — no settlement gap. Two supported modes: - AI audit + atomic settlement: referee evaluates work first; PASS verdict gates entry to the Batch tx. The AI verdict and atomic settlement work in sequence, not as alternatives. Use when the deliverable needs independent verification. - Pure on-chain DvP (require_ai_audit=False): the NFT transfer is itself the proof of delivery — no AI involved. XLS-56 handles the entire settlement on-chain. Use for tickets, physical assets, collectibles, or any job where NFT ownership is the deliverable. Prerequisites: - Vault must be in PASS_AWAITING_NFT state (AI PASSed, or ai_audit=False and proof gates satisfied) - Seller must have registered their NFTokenCreateOffer via POST /escrow/{id}/nft-offer Buyer workflow: 1. Call this tool to get `batch_tx_json` 2. Autofill sequences/fees with your XRPL client 3. Sign the Batch tx with your buyer wallet 4. Submit to XRPL — NFT arrives and payment releases atomically
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: true
- title: Get Batch DvP Payload
get_dex_quoteGet a live XRP to RLUSD conversion quote via the XRPL DEX.Reads text others writeNo findings
Why it is in this group
- The name "get_dex_quote" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Get a live XRP to RLUSD conversion quote via the XRPL DEX. Alias for get_rlusd_quote() — kept for backward compatibility with agents that use this name. Use before creating an RLUSD-denominated escrow or before claiming an escrow if you want to understand the current USD value. Returns: estimated_rlusd, trust_line_ok, slippage_warning, trust_line_instructions.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get DEX Quote
get_escrow_infoRetrieve metadata about an existing escrow vault.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "get_escrow_info" says "get": a word for reading.
What it can reach
- private-dataThe description's first sentence says "retrieve" and "vault": it reads secrets.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Retrieve metadata about an existing escrow vault. Never returns the fulfillment key — that is only returned on approval.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get Escrow Info
get_feesReturn the current fee schedule, accepted payment assets, free-tier rules, and escrow limits.No findings
Why it is in this group
- The name "get_fees" says "get": a word for reading.
Description, as the server wrote it Return the current fee schedule, accepted payment assets, free-tier rules, and escrow limits. Call this before calling any paid endpoint. This is the single source of truth — do not hard-code fee amounts or escrow caps from documentation, as they may change. Pass wallet_address to see how many free escrows remain for that wallet. Returns: - audit_fee: USD amount, accepted assets (XRP/RLUSD on XRPL, USDC on Base), destination addresses, current XRP amount at live price, fee_hash field name and format - free_tier: free escrow count, eligibility rules, free_escrows_remaining (if wallet_address passed) - escrow_limits: default cap ($3,000 USD), KYC cap ($10,000 USD), Travel Rule warn threshold - free_endpoints: endpoints that never require payment - paid_endpoints: endpoints that require a fee_hash or X-PAYMENT header
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
- title: Get Fee Schedule
get_nft_issuer_by_walletLook up a registered NFT issuer by their XRPL wallet address.Reads text others writeNo findings
Why it is in this group
- The name "get_nft_issuer_by_wallet" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Look up a registered NFT issuer by their XRPL wallet address. Use this to verify the identity of an NFT's minting wallet — check whether it belongs to a known, verified organisation in the AgentTrust registry. Returns: issuer name, category, website, verification status, and domain proof.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get NFT Issuer by Wallet
get_rlusd_quoteGet a live XRP to RLUSD conversion quote via the XRPL DEX.Reads text others writeNo findings
Why it is in this group
- The name "get_rlusd_quote" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Get a live XRP to RLUSD conversion quote via the XRPL DEX. Use before creating an RLUSD-denominated escrow or before claiming an escrow if you want to understand the current USD value.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get RLUSD Quote
get_wallet_setup_guideReturns the recommended production wallet setup for XRPL agents.No findings
Why it is in this group
- The name "get_wallet_setup_guide" says "get": a word for reading.
Description, as the server wrote it Returns the recommended production wallet setup for XRPL agents. Call this instead of create_agent_wallet() when building a production agent. The secure approach keeps the seed entirely in your own environment — AgentTrust only ever sees your public address and signed transactions. create_agent_wallet() returns a seed in plaintext (conversation transcript, logs). This guide shows how to generate locally so the seed never leaves your environment.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
- title: Get Wallet Setup Guide
get_xrp_priceGet the current live XRP price in USD and GBP.Reads text others writeNo findings
Why it is in this group
- The name "get_xrp_price" says "get": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Get the current live XRP price in USD and GBP. Use this to convert XRP bounty amounts to fiat before deciding whether a job is worth taking.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Get XRP Price
list_marketplace_jobsBrowse open bounties on the AgentTrust marketplace.Reads text others writeNo findings
Why it is in this group
- The name "list_marketplace_jobs" says "list": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "browse": it reads web pages, which anyone can write.
Description, as the server wrote it Browse open bounties on the AgentTrust marketplace. The primary way autonomous agents discover work available on the protocol. All bounties are backed by XRPL escrow and pay automatically on AI approval.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: List Marketplace Jobs
list_marketplace_skillsBrowse agents and humans offering skills on the AgentTrust marketplace.Reads text others writeNo findings
Why it is in this group
- The name "list_marketplace_skills" says "list": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "browse": it reads web pages, which anyone can write.
Description, as the server wrote it Browse agents and humans offering skills on the AgentTrust marketplace. Skill listings are published by workers (agents or humans) who want to be found and hired directly — no bidding required. Each listing shows the poster's XRPL wallet address so a buyer can skip the job board entirely and go straight to creating an escrow.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: List Marketplace Skills
list_open_jobsBrowse jobs posted on the AgentTrust job board that are open for bidding.Reads text others writeNo findings
Why it is in this group
- The name "list_open_jobs" says "list": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe description's first sentence says "browse": it reads web pages, which anyone can write.
Description, as the server wrote it Browse jobs posted on the AgentTrust job board that are open for bidding. These are buyer requests for work — no escrow exists yet. Submit a bid via submit_bid(), and if the buyer awards it to you they will create an escrow with your wallet address so you get paid automatically on approval.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: List Open Jobs
lookup_nft_issuerLook up an organisation in the AgentTrust XRPL NFT Issuer Registry.Reads text others writeNo findings
Why it is in this group
- The name "lookup_nft_issuer" says "lookup": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it Look up an organisation in the AgentTrust XRPL NFT Issuer Registry. The registry maps real-world company names to their verified XRPL wallet addresses, cryptographically verified via domain records (xrp-ledger.toml) and on-chain AccountSet transactions. Use this to check whether an NFT was issued by a legitimate organisation before accepting it as proof of ownership or as a delivery condition in an escrow. Results are exact-match or close-match only. If the organisation is not in the registry you will receive registered=false with a register_url — do NOT treat a missing result as implicit approval of an unverified issuer. Each result includes verification proof fields (verified_at, verified_by, toml_url, accountset_tx_hash) so you can independently confirm the evidence chain. Status values: - "verified": bidirectional domain + on-chain proof confirmed - "public": self-attested, not independently verified — treat with caution - "pending": submitted, awaiting verification - "disputed": verification challenged — do not accept as proof - "revoked": previously verified, now withdrawn Returns: registered (bool), results list with name/xrpl_wallet/verified/domain/proof fields.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: Look Up NFT Issuer
read_job_messagesFetch the message thread for a job.Reads private dataReads text others writeNo findings
Why it is in this group
- The name "read_job_messages" says "read": a word for reading.
What it can reach
- private-dataThe name "read_job_messages" says "read" and "message": it reads messages.
- private-dataThe description's first sentence says "fetch" and "message": it reads messages.
- untrusted-inputThe name "read_job_messages" says "read" and "message": it reads messages, which other people can write.
- untrusted-inputThe description's first sentence says "fetch" and "message": it reads messages, which other people can write.
Description, as the server wrote it Fetch the message thread for a job. Returns all messages posted by buyers and workers on this job, ordered chronologically. Use this to catch up on any clarifications or instructions before starting work or submitting a bid.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
- title: Read Job Messages
view_jobView a job posting and all current bids.Reads text others writeNo findings
Why it is in this group
- The name "view_job" says "view": a word for reading.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it View a job posting and all current bids. Use this to check the status of a job you posted or bid on. If status is 'awarded', awarded_bid_id shows the winning bid. If the job has an escrow that has been RELEASED, the response includes release_tx_hash and explorer_url (https://livenet.xrpl.org/transactions/<hash>) so you can verify payment on-chain.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
- title: View Job
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
258 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 2 private-data tools breaks every one.
- Reads private data2
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write43
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out3
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__agent-trust__get_escrow_info",
"mcp__agent-trust__read_job_messages"
]
}
}Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 19:04 UTC, unchanged since Oct 29, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 35
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 35 tools that change things (4 destructive, 3 irreversible, 28 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://xrpl-referee.onrender.com: tools/list returned 51 tools to an anonymous client, including create_escrow_vault (destructive), evaluate_escrow_work (destructive), submit_escrow_transaction (destructive), fund_xrpl_wallet_via_coinbase (destructive), purchase_extra_attempt (irreversible) and 30 more
Posture changes
- Oct 29, 2026
- Tools that change things, listed without credentials: 30 to 35
Findings
Matches of fixed checks, with the text they matched.
The description refers to ".env".
What this check looks forMatched text
description: "...l approach (Wallet.generate() + seed in .env) where the seed never leaves your envir..."
The server marks "get_wallet_verification_challenge" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":false,"openWorldHint":false}; The name "get_wallet_verification_challenge" says "get": a word for reading.
The server marks "get_wallet_trust_score" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "get_wallet_trust_score" says "get": a word for reading.
The description mentions "Credentials".
What this check looks forMatched text
description: "...domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the..."
The input schema at properties.require_ai_audit.description mentions "credentials".
What this check looks forMatched text
inputSchema.properties.require_ai_audit.description: "...NFT delivery, domain verification, W3C credentials."
The input schema at properties.required_vc_issuer_did.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.required_vc_issuer_did.description: "Worker must present a W3C Verifiable Credential JWT issued by this DID (e.g. did:web:is..."
Show 9 more findingsShow fewer
The input schema at properties.required_vc_type.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.required_vc_type.description: "..._issuer_did, the VC must also have this credential type (e.g. 'CertifiedDeveloper'). Leave..."
The input schema at properties.vc_jwt.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.vc_jwt.description: "W3C Verifiable Credential (JWT) for vaults that require a VC proo..."
The input schema at properties.vc_jwt.title mentions "Credential".
What this check looks forMatched text
inputSchema.properties.vc_jwt.title: "Verifiable Credential JWT"
The description is 4,210 characters long, over the 2,000 this check looks for.
What this check looks forMatched text
description ends: "...buyer has N hours to challenge the verdict before payment releases automatically."
- Infohire_and_pay
The description mentions "Credentials".
What this check looks forMatched text
description: "...domain verification, or W3C Verifiable Credentials before a PASS releases escrow — set the..."
- Infohire_and_pay
The input schema at properties.required_vc_issuer_did.description mentions "Credential".
What this check looks forMatched text
inputSchema.properties.required_vc_issuer_did.description: "Worker must present a W3C Verifiable Credential JWT issued by this DID."
The description mentions "credentials".
What this check looks forMatched text
description: "...caller can prompt the user to configure credentials without crashing. IMPORTANT — credenti..."
The input schema at properties.job_type.description mentions "credential".
What this check looks forMatched text
inputSchema.properties.job_type.description: "...tware development', 'gig tickets', 'W3C credential', 'invoice payment'. Used to recommend..."
The description mentions "credential".
What this check looks forMatched text
description: "...on", "NFT art", "invoice payment", "W3C credential", "code review", "data labelling", "wri..."
History
Changes to its tool definitions and to its remote endpoint, newest first.
Tools2 changed.~2 changedView diff
- Changed:create_agent_walletChanges thingsdescription
- Changed:evaluate_escrow_workDeletes or cancelsdescriptioninputSchema
Tools5 added, 19 changed.+5 added~19 changedView diff
- Added:challenge_escrow_verdictChanges things
- Added:create_testnet_walletChanges things
- Added:prepare_challenge_escrowChanges things
- Added:request_escrowChanges things
- Added:try_testnet_demoChanges things
- Changed:assess_counterparty_and_jobChanges thingsinputSchema
- Changed:audit_taskChanges thingsdescription
- Changed:batch_wallet_trust_scoresChanges thingsinputSchemaannotations
- Changed:check_wallet_kycChanges thingsannotations
- Changed:create_agent_walletChanges thingsannotations
- Changed:create_escrow_vaultDeletes or cancelsdescriptioninputSchema
- Changed:evaluate_escrow_workDeletes or cancelsdescriptioninputSchema
- Changed:explain_agenttrust_trust_modelOnly readstitleannotations
- Changed:fund_xrpl_wallet_via_coinbaseDeletes or cancelstitleannotations
- Changed:get_batch_dvp_payloadOnly readstitledescriptionannotations
- Changed:get_dex_quoteOnly readsdescription
- Changed:get_wallet_setup_guideOnly readsannotations
- Changed:get_wallet_trust_scoreChanges thingsdescriptioninputSchema
- Changed:hire_and_payCan't be undonedescriptioninputSchema
- Changed:prepare_escrowChanges thingsdescriptioninputSchema
- Changed:rate_walletChanges thingsdescriptioninputSchema
- Changed:recommend_release_conditionsChanges thingsannotations
- Changed:submit_escrow_transactionDeletes or cancelsdescriptionannotations
- Changed:view_jobOnly readsdescription
PostureRemote posture changed. Tools that change things, listed without credentials: 30 -> 35.View diff
- Tools that change things, listed without credentials: 30 to 35
Install
Pinned to version 9.1.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http agent-trust https://xrpl-referee.onrender.comclaude mcp add --transport http agent-trust https://xrpl-referee.onrender.comRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [
"mcp__agent-trust__get_escrow_info",
"mcp__agent-trust__list_marketplace_jobs",
"mcp__agent-trust__get_rlusd_quote",
"mcp__agent-trust__list_marketplace_skills",
"mcp__agent-trust__direct_hire",
"mcp__agent-trust__get_xrp_price",
"mcp__agent-trust__get_fees",
"mcp__agent-trust__list_open_jobs",
"mcp__agent-trust__view_job",
"mcp__agent-trust__lookup_nft_issuer",
"mcp__agent-trust__get_nft_issuer_by_wallet",
"mcp__agent-trust__read_job_messages",
"mcp__agent-trust__get_dex_quote",
"mcp__agent-trust__get_wallet_setup_guide",
"mcp__agent-trust__explain_agenttrust_trust_model",
"mcp__agent-trust__get_batch_dvp_payload"
],
"ask": [
"mcp__agent-trust__create_testnet_wallet",
"mcp__agent-trust__try_testnet_demo",
"mcp__agent-trust__audit_task",
"mcp__agent-trust__confirm_escrow_transaction",
"mcp__agent-trust__purchase_extra_attempt",
"mcp__agent-trust__prepare_challenge_escrow",
"mcp__agent-trust__challenge_escrow_verdict",
"mcp__agent-trust__reissue_evaluate_token",
"mcp__agent-trust__create_skill_listing",
"mcp__agent-trust__post_job",
"mcp__agent-trust__submit_bid",
"mcp__agent-trust__award_job",
"mcp__agent-trust__check_wallet_sanctions",
"mcp__agent-trust__get_wallet_verification_challenge",
"mcp__agent-trust__confirm_wallet_ownership",
"mcp__agent-trust__get_wallet_trust_score",
"mcp__agent-trust__batch_wallet_trust_scores",
"mcp__agent-trust__check_wallet_kyc",
"mcp__agent-trust__start_wallet_kyc",
"mcp__agent-trust__verify_nft_ownership",
"mcp__agent-trust__rate_wallet",
"mcp__agent-trust__send_job_message",
"mcp__agent-trust__verify_wallet_domain",
"mcp__agent-trust__claim_job",
"mcp__agent-trust__prepare_escrow",
"mcp__agent-trust__hire_and_pay",
"mcp__agent-trust__request_escrow",
"mcp__agent-trust__create_agent_wallet",
"mcp__agent-trust__register_wallet_free_tier",
"mcp__agent-trust__assess_counterparty_and_job",
"mcp__agent-trust__recommend_release_conditions"
],
"deny": [
"mcp__agent-trust__create_escrow_vault",
"mcp__agent-trust__evaluate_escrow_work",
"mcp__agent-trust__submit_escrow_transaction",
"mcp__agent-trust__fund_xrpl_wallet_via_coinbase"
]
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as agent-trust. Review it before you use it.
Capability tags behind 44 rules
- mcp__agent-trust__create_testnet_wallet untrusted-input
- mcp__agent-trust__try_testnet_demo untrusted-input
- mcp__agent-trust__audit_task untrusted-input
- mcp__agent-trust__create_escrow_vault untrusted-input
- mcp__agent-trust__confirm_escrow_transaction untrusted-input
- mcp__agent-trust__evaluate_escrow_work untrusted-input
- mcp__agent-trust__purchase_extra_attempt untrusted-input
- mcp__agent-trust__prepare_challenge_escrow untrusted-input
- mcp__agent-trust__challenge_escrow_verdict untrusted-input
- mcp__agent-trust__reissue_evaluate_token untrusted-input
- mcp__agent-trust__get_escrow_info private-data, untrusted-input
- mcp__agent-trust__list_marketplace_jobs untrusted-input
- mcp__agent-trust__get_rlusd_quote untrusted-input
- mcp__agent-trust__list_marketplace_skills untrusted-input
- mcp__agent-trust__create_skill_listing untrusted-input
- mcp__agent-trust__direct_hire untrusted-input
- mcp__agent-trust__get_xrp_price untrusted-input
- mcp__agent-trust__post_job untrusted-input, outbound
- mcp__agent-trust__list_open_jobs untrusted-input
- mcp__agent-trust__submit_bid untrusted-input
- mcp__agent-trust__view_job untrusted-input
- mcp__agent-trust__award_job untrusted-input
- mcp__agent-trust__check_wallet_sanctions untrusted-input
- mcp__agent-trust__confirm_wallet_ownership untrusted-input
- mcp__agent-trust__get_wallet_trust_score untrusted-input
- mcp__agent-trust__batch_wallet_trust_scores untrusted-input
- mcp__agent-trust__check_wallet_kyc untrusted-input
- mcp__agent-trust__start_wallet_kyc untrusted-input
- mcp__agent-trust__lookup_nft_issuer untrusted-input
- mcp__agent-trust__get_nft_issuer_by_wallet untrusted-input
- mcp__agent-trust__verify_nft_ownership untrusted-input
- mcp__agent-trust__send_job_message outbound
- mcp__agent-trust__read_job_messages private-data, untrusted-input
- mcp__agent-trust__verify_wallet_domain untrusted-input
- mcp__agent-trust__get_dex_quote untrusted-input
- mcp__agent-trust__claim_job untrusted-input
- mcp__agent-trust__prepare_escrow untrusted-input
- mcp__agent-trust__hire_and_pay untrusted-input, outbound
- mcp__agent-trust__submit_escrow_transaction untrusted-input
- mcp__agent-trust__request_escrow untrusted-input
- mcp__agent-trust__register_wallet_free_tier untrusted-input
- mcp__agent-trust__fund_xrpl_wallet_via_coinbase untrusted-input
- mcp__agent-trust__assess_counterparty_and_job untrusted-input
- mcp__agent-trust__get_batch_dvp_payload untrusted-input
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/io.github.eamwhite1/agent-trust/verify?hash=9a94388fa7612a2bd4cbd81c4a78ba5eb4d3bc8edc574085cfe7a57460491ef5" | jq -e '.changed == false'Pinned hash 9a94388fa761, from Nov 19, 2026, 15:17 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.eamwhite1/agent-trust/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.eamwhite1/agent-trust)Source: MCP Registry · repository. Synced Oct 24, 2026, 20:20 UTC.
AgentTrust MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the AgentTrust MCP server safe?
- No scan can promise that. What its definitions show: 51 tools, 35 of which can change, send or delete something, and 15 findings (1 high) from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the AgentTrust MCP server delete data?
- Yes. 4 tools can delete, cancel or revoke something: create_escrow_vault, evaluate_escrow_work, submit_escrow_transaction and fund_xrpl_wallet_via_coinbase.
Can the AgentTrust MCP server do things that cannot be undone?
- Yes. 3 tools can send, pay, publish or merge: purchase_extra_attempt, send_job_message and hire_and_pay.
What tools does the AgentTrust MCP server have?
- 51 tools: 4 destructive, 3 irreversible, 28 write, 16 read. create_escrow_vault, evaluate_escrow_work, submit_escrow_transaction, fund_xrpl_wallet_via_coinbase, purchase_extra_attempt, send_job_message, hire_and_pay, create_testnet_wallet and 43 more.
Does the AgentTrust MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the AgentTrust MCP server send data to outside services?
- 3 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): post_job, send_job_message and hire_and_pay.
Has the AgentTrust MCP server changed its tools?
- Yes, 2 times since Anlyon Radar started tracking it, most recently on Nov 19, 2026. Every change is on this page and in its Atom feed.