Skip to content

agentforge MCP server

io.github.acarchidi/agentforgeacarchidistreamable-httpv1.13.1Scanned

Safety checks for AI agents: preflight before signing, prompt-injection scan, sanctions, rug checks.

29 tools: 2 do something that can't be undone, 20 change things and 7 only read.

Tools: 0 delete, 2 irreversible, 20 write, 7 read.
  • Reads private data · 0
  • Reads text others write · 29
  • Can send data out · 2
  • 3 findings

Tools listed without sign-in. Remote posture

With Anlyon

2 tools here do something that can't be undone. Govern them.

  • Anlyon lists the tools and adopts none.
  • Each tool you adopt gets approvals and receipts.
  • Your agent never holds the key.
Govern this server →

A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.

What the agentforge MCP server can do

The agentforge MCP server is published by the GitHub account acarchidi. It has 29 tools. It can do things that cannot be undone, such as sending, paying or publishing (preflight and send_feedback) and create or change data (20 tools: token_intel, token_research, contract_docs, contract_monitor and 16 more), but cannot delete data. 7 tools only read.

29 tools bring in text other people can write (token_intel, token_research, contract_docs and 26 more); 2 tools can send data out of your system (preflight and send_feedback).

The fixed checks matched 3 findings (annotation-conflict). Each one comes with the text it matched.

Its remote endpoint lists its tools without sign-in, and its tools changed 5 times since Anlyon Radar first saw them.

Last scanned Nov 9, 2026, version 1.13.1.

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.

29 tools

Can't be undone 2Claude Code: ask
  • preflightIs this action safe?Reads text others writeCan send data out1 finding

    Why it is in this group

    • The input addresses people (input.recipients): what it sends cannot be taken back.
    • 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.
    • outboundThe input takes a destination (input.recipients): what the tool writes goes there.

    Findings

    • Warn

      The server marks "preflight" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The input addresses people (input.recipients): what it sends cannot be taken back.

      What this check looks for
    Description, as the server wrote it
    Is this action safe? Call once before any on-chain action: send the transactions, signature request, x402 payment request, recipients, and/or tokens the agent is about to act on, and get back one proceed/caution/avoid verdict. Transactions are simulated (reverts, assets out, approvals, counterparty screening), signatures decoded and checked for permit phishing, recipients screened against sanctions and scam lists, and tokens rug-checked, including tokens the simulation shows the wallet receiving. Returns one proceed / caution / avoid verdict with reasons. Ethereum, Base, Arbitrum, Optimism, Polygon, BNB Chain, Avalanche, Linea, Unichain, Sonic and HyperEVM. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Preflight
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • send_feedbackSend AgentForge feedback: a bug or wrong result, a complaint, a feature or new-check request, or an idea.Reads text others writeCan send data out1 finding

    Why it is in this group

    • The name "send_feedback" says "send": a word for sending, and what is sent cannot be taken back.
    • 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.
    • outboundThe name "send_feedback" says "send": it sends or publishes something to other people.
    • outboundThe input takes a URL (input.endpoint): a request to a URL can carry data to whoever runs it.

    Findings

    • Warn

      The server marks "send_feedback" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "send_feedback" says "send": a word for sending, and what is sent cannot be taken back.

      What this check looks for
    Description, as the server wrote it
    Send AgentForge feedback: a bug or wrong result, a complaint, a feature or new-check request, or an idea. Free. Include a contact (email, wallet, or URL) if you want a reply.

    Annotations the server declares

    • title: Send Feedback
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Changes things 20Claude Code: ask
  • address_checkScreen counterparty addresses before sending, approving, or interacting: OFAC sanctions, known scam/phishing wallets, labeled protocol contracts, and brand-new-address detection.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 counterparty addresses before sending, approving, or interacting: OFAC sanctions, known scam/phishing wallets, labeled protocol contracts, and brand-new-address detection. 1-100 EVM or Solana addresses per call; returns block / caution / no_known_issues per address. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Address Check
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • approval_scanScan a wallet for risky token approvals.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
    Scan a wallet for risky token approvals. On EVM chains: identifies unlimited ERC-20 approvals and unverified spenders. On Solana: scans SPL token delegate authorities. Returns risk assessment. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Approval Scan
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • content_scanScan untrusted text (tool results, web pages, emails, token metadata) for prompt injection before feeding it to your LLM: instruction overrides, fund-movement directives, payloads …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
    Scan untrusted text (tool results, web pages, emails, token metadata) for prompt injection before feeding it to your LLM: instruction overrides, fund-movement directives, payloads hidden in base64, Morse, zero-width or tag characters and homoglyphs, leaked private keys and seed phrases, embedded addresses screened against sanctions and scam lists, lookalike domains. Deterministic, no LLM. Returns clean / suspicious / malicious with findings. Up to 64 KB per call. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Content Scan
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • contract_docsGenerate documentation for any verified EVM smart contract.Reads text others write1 finding

    Why it is in this group

    • The description's first verb is "generate": 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 "contract_docs" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

      Matched text

      annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The description's first verb is "generate": a word for a change.

      What this check looks for
    Description, as the server wrote it
    Generate documentation for any verified EVM smart contract. Returns function descriptions, risk flags, interaction patterns, and security posture. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Contract Docs
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • contract_monitorMonitor recent contract activity for suspicious admin operations, proxy upgrades, ownership changes, and pause events.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
    Monitor recent contract activity for suspicious admin operations, proxy upgrades, ownership changes, and pause events. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Contract Monitor
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • gas_oracleCurrent gas prices (slow/standard/fast) for any supported EVM chain with trend analysis.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
    Current gas prices (slow/standard/fast) for any supported EVM chain with trend analysis.

    Annotations the server declares

    • title: Gas Oracle
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • launch_feedNewly launched tokens on Base, Ethereum, or Arbitrum, each already risk-scored (verification, owner powers, proxy, supply held off-pool, thin liquidity, honeypot trading pattern).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
    Newly launched tokens on Base, Ethereum, or Arbitrum, each already risk-scored (verification, owner powers, proxy, supply held off-pool, thin liquidity, honeypot trading pattern). Poll it: pass the previous nextSince as since to get only new launches. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Launch Feed
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • lending_healthHow close a wallet is to liquidation on Aave v3 (Ethereum, Base, Arbitrum, Optimism, Polygon, BNB Chain, Avalanche, Linea, and Sonic in one call): collateral, debt, health factor, …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
    How close a wallet is to liquidation on Aave v3 (Ethereum, Base, Arbitrum, Optimism, Polygon, BNB Chain, Avalanche, Linea, and Sonic in one call): collateral, debt, health factor, status, and how far collateral can fall before liquidation. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Lending Health
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • risk_batchBatch rug-check: risk scores (holder concentration, mint/blacklist/pause/upgrade permissions, liquidity depth) for 2-25 EVM tokens in one call.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
    Batch rug-check: risk scores (holder concentration, mint/blacklist/pause/upgrade permissions, liquidity depth) for 2-25 EVM tokens in one call. Use it to screen a watchlist or a list of candidate trades at once. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Risk Batch
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • signature_checkCheck a signature request BEFORE signing it: decodes EIP-712 typed data (ERC-20 permit, Permit2, Seaport order, transfer authorization) or an x402 payment request, explains in one …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 a signature request BEFORE signing it: decodes EIP-712 typed data (ERC-20 permit, Permit2, Seaport order, transfer authorization) or an x402 payment request, explains in one sentence what signing would allow, screens spenders and payees for sanctions and scams, and catches signature-drainer patterns. Returns proceed / caution / avoid. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Signature Check
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • solana_token_risk_scanSolana token rug check: mint/freeze authority, holder concentration, liquidity depth, and a composite 0-100 risk score.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
    Solana token rug check: mint/freeze authority, holder concentration, liquidity depth, and a composite 0-100 risk score. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Solana Token Risk Scan
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • solana_tx_simulateSimulate a Solana transaction before signing it: balance changes, labeled programs, deterministic risk rules, and a proceed/caution/avoid recommendation.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
    Simulate a Solana transaction before signing it: balance changes, labeled programs, deterministic risk rules, and a proceed/caution/avoid recommendation. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Solana Tx Simulate
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • token_intelLightweight token lookup: price, market cap, volume, and basic risk assessment for any EVM or Solana token.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
    Lightweight token lookup: price, market cap, volume, and basic risk assessment for any EVM or Solana token. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Token Intel
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • token_researchMulti-source token intelligence: market data, DeFi metrics, contract verification, prediction markets, holder analysis, price history, and AI risk assessment.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
    Multi-source token intelligence: market data, DeFi metrics, contract verification, prediction markets, holder analysis, price history, and AI risk assessment. Aggregates CoinGecko, DeFiLlama, Etherscan, and Polymarket. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Token Research
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • token_risk_metricsQuantitative risk metrics for any ERC-20 token: holder concentration (top 10 holder %), contract permissions (can mint/burn/pause/blacklist), liquidity depth vs market cap, deploye…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
    Quantitative risk metrics for any ERC-20 token: holder concentration (top 10 holder %), contract permissions (can mint/burn/pause/blacklist), liquidity depth vs market cap, deployer history, and weighted composite risk score (0-100). Pre-computed for top tokens, live-computed for others. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Token Risk Metrics
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • trade_checkCheck a token before you buy it: one call bundling quantitative rug-check metrics (holder concentration, permissions, liquidity), AI risk assessment with market data, and current g…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 a token before you buy it: one call bundling quantitative rug-check metrics (holder concentration, permissions, liquidity), AI risk assessment with market data, and current gas — rolled into a proceed/caution/avoid verdict with reasons. Cheaper than calling the three services separately. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Trade Check
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • translateTranslate text to any language with tone control (formal/casual/technical).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
    Translate text to any language with tone control (formal/casual/technical). Auto-detects source language. Preserves formatting and cultural nuances. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Translate
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • tx_decodeDecode any EVM transaction: function call, parameters, token transfers, and plain-English explanation.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
    Decode any EVM transaction: function call, parameters, token transfers, and plain-English explanation. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Tx Decode
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • tx_simulateSimulate an EVM transaction (or a sequence of up to 5, e.g.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
    Simulate an EVM transaction (or a sequence of up to 5, e.g. approve then swap) before signing: revert check with reason, every asset leaving or entering the wallet, approvals granted, sanctions/scam screening of counterparties, and a proceed/caution/avoid verdict. Ethereum, Base, Arbitrum, Optimism, Polygon, BNB Chain, Linea, Unichain, Sonic. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Tx Simulate
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • wallet_safetyComprehensive wallet safety check: scans approvals, analyzes recent transaction activity for suspicious patterns, and assesses target contract risk.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
    Comprehensive wallet safety check: scans approvals, analyzes recent transaction activity for suspicious patterns, and assesses target contract risk. Returns composite risk score (0-100), risk level, action items, and related service suggestions. Supports EVM chains and Solana. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Wallet Safety
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
Only reads 7Show 7 read-only tools
  • pool_snapshotGet a cached snapshot of top DeFi liquidity pools.Reads text others writeNo findings

    Why it is in this group

    • The name "pool_snapshot" says "snapshot": 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 cached snapshot of top DeFi liquidity pools. Filter by protocol (e.g. "uniswap-v3"), chain (e.g. "ethereum"), or token symbol (e.g. "ETH"). Returns TVL, APY, 24h volume, IL risk, and registry enrichment. Data refreshed every 15 minutes.

    Annotations the server declares

    • title: Pool Snapshot
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • registry_lookupLook up a contract address in the Known Contract Label Registry.Reads text others writeNo findings

    Why it is in this group

    • The name "registry_lookup" 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 a contract address in the Known Contract Label Registry. Returns protocol name, category, risk level, and tags. Free — no payment required.

    Annotations the server declares

    • title: Registry Lookup
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • sentimentAnalyze sentiment of text in crypto, finance, social media, or general context.Reads text others writeNo findings

    Why it is in this group

    • The description's first verb is "analyze": 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
    Analyze sentiment of text in crypto, finance, social media, or general context. Returns score (-1 to 1), confidence, label (very_bearish to very_bullish), reasoning, and per-entity sentiment. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Sentiment
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • solana_program_lookupLook up a Solana program ID in the program label registry.Reads text others writeNo findings

    Why it is in this group

    • The name "solana_program_lookup" 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 a Solana program ID in the program label registry. Returns protocol name, category, risk level. Free — no payment required.

    Annotations the server declares

    • title: Solana Program Lookup
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • solana_tx_explainExplain a Solana transaction in plain English: labeled programs, token/SOL movements, success status, and risk flags.Reads text others writeNo findings

    Why it is in this group

    • The name "solana_tx_explain" says "explain": 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
    Explain a Solana transaction in plain English: labeled programs, token/SOL movements, success status, and risk flags. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Solana Tx Explain
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • summarizeSummarize text with configurable length (brief/standard/detailed), format (prose/bullet_points/structured), and optional topic focus.Reads text others writeNo findings

    Why it is in this group

    • The name "summarize" says "summarize": 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
    Summarize text with configurable length (brief/standard/detailed), format (prose/bullet_points/structured), and optional topic focus. Returns summary, key points, and compression ratio. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Summarize
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true
  • token_compareCompare a primary token against up to 3 others.Reads text others writeNo findings

    Why it is in this group

    • The name "token_compare" says "compare": 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
    Compare a primary token against up to 3 others. Returns full research on primary, abbreviated metrics on comparisons, plus AI comparative analysis. Free over MCP for 10 calls per day; beyond that the tool returns the pay-per-call x402 endpoint and its price.

    Annotations the server declares

    • title: Token Compare
    • readOnlyHint: true
    • destructiveHint: false
    • idempotentHint: true
    • openWorldHint: true

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.

No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.

Reads private data0
Reads files, mail, messages, databases, secrets, calendars or repositories.
Reads text others write29
Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
Can send data out2
Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.

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.

Tools listed without sign-in

Looked at Nov 20, 2026, 19:01 UTC, unchanged since Nov 6, 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
22
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 22 tools that change things (2 irreversible, 20 write). Whether a call also works without credentials was not tested: tools are never called.

    Matched text

    https://agentforge-taupe.vercel.app/mcp: tools/list returned 29 tools to an anonymous client, including preflight (irreversible), send_feedback (irreversible), token_intel (write), token_research (write), contract_docs (write) and 17 more

    What this check looks for

Posture changes

  1. Nov 6, 2026
    • Tools that change things, listed without credentials: 13 to 22

Findings

Matches of fixed checks, with the text they matched.

3 warn
  • The server marks "contract_docs" read-only (readOnlyHint: true), and its own name, description or input give the class "write".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The description's first verb is "generate": a word for a change.

    What this check looks for
  • The server marks "preflight" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The input addresses people (input.recipients): what it sends cannot be taken back.

    What this check looks for
  • The server marks "send_feedback" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".

    Matched text

    annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "send_feedback" says "send": a word for sending, and what is sent cannot be taken back.

    What this check looks for

History

Changes to its tool definitions and to its remote endpoint, newest first.

  1. PostureRemote posture changed. Tools that change things, listed without credentials: 13 -> 22.View diff
    • Tools that change things, listed without credentials: 13 to 22
  2. Tools1 added.+1 addedView diff
  3. Tools1 added, 8 changed.+1 added~8 changedView diff
  4. Tools1 added, 9 changed.+1 added~9 changedView diff
  5. Tools3 added, 1 removed, 23 changed.+3 added−1 removed~23 changedView diff
  6. Tools4 added.+4 addedView diff

Install

Pinned to version 1.13.1. Values in angle brackets are yours to fill in.

claude mcp add --transport http agentforge https://agentforge-taupe.vercel.app/mcp
claude mcp add --transport http agentforge https://agentforge-taupe.vercel.app/mcp

Run 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.

.claude/settings.json
{
  "permissions": {
    "allow": [
      "mcp__agentforge__token_compare",
      "mcp__agentforge__sentiment",
      "mcp__agentforge__summarize",
      "mcp__agentforge__pool_snapshot",
      "mcp__agentforge__registry_lookup",
      "mcp__agentforge__solana_program_lookup",
      "mcp__agentforge__solana_tx_explain"
    ],
    "ask": [
      "mcp__agentforge__token_intel",
      "mcp__agentforge__token_research",
      "mcp__agentforge__contract_docs",
      "mcp__agentforge__contract_monitor",
      "mcp__agentforge__tx_decode",
      "mcp__agentforge__approval_scan",
      "mcp__agentforge__gas_oracle",
      "mcp__agentforge__translate",
      "mcp__agentforge__wallet_safety",
      "mcp__agentforge__token_risk_metrics",
      "mcp__agentforge__trade_check",
      "mcp__agentforge__risk_batch",
      "mcp__agentforge__address_check",
      "mcp__agentforge__content_scan",
      "mcp__agentforge__launch_feed",
      "mcp__agentforge__preflight",
      "mcp__agentforge__tx_simulate",
      "mcp__agentforge__signature_check",
      "mcp__agentforge__lending_health",
      "mcp__agentforge__send_feedback",
      "mcp__agentforge__solana_tx_simulate",
      "mcp__agentforge__solana_token_risk_scan"
    ],
    "deny": []
  }
}

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 agentforge. Review it before you use it.

Capability tags behind 29 rules
  • mcp__agentforge__token_intel untrusted-input
  • mcp__agentforge__token_research untrusted-input
  • mcp__agentforge__contract_docs untrusted-input
  • mcp__agentforge__contract_monitor untrusted-input
  • mcp__agentforge__token_compare untrusted-input
  • mcp__agentforge__tx_decode untrusted-input
  • mcp__agentforge__approval_scan untrusted-input
  • mcp__agentforge__gas_oracle untrusted-input
  • mcp__agentforge__sentiment untrusted-input
  • mcp__agentforge__summarize untrusted-input
  • mcp__agentforge__translate untrusted-input
  • mcp__agentforge__wallet_safety untrusted-input
  • mcp__agentforge__pool_snapshot untrusted-input
  • mcp__agentforge__token_risk_metrics untrusted-input
  • mcp__agentforge__trade_check untrusted-input
  • mcp__agentforge__risk_batch untrusted-input
  • mcp__agentforge__address_check untrusted-input
  • mcp__agentforge__content_scan untrusted-input
  • mcp__agentforge__launch_feed untrusted-input
  • mcp__agentforge__preflight untrusted-input, outbound
  • mcp__agentforge__tx_simulate untrusted-input
  • mcp__agentforge__signature_check untrusted-input
  • mcp__agentforge__lending_health untrusted-input
  • mcp__agentforge__send_feedback untrusted-input, outbound
  • mcp__agentforge__registry_lookup untrusted-input
  • mcp__agentforge__solana_program_lookup untrusted-input
  • mcp__agentforge__solana_tx_explain untrusted-input
  • mcp__agentforge__solana_tx_simulate untrusted-input
  • mcp__agentforge__solana_token_risk_scan 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.acarchidi/agentforge/verify?hash=71089500ab8ecba718e37ed250e91fecdfb2a3a97397770f5760527fffa5e215" | jq -e '.changed == false'

Pinned hash 71089500ab8e, from Nov 9, 2026, 21:18 UTC.

Watch

Get told when its tools change: a new tool, a new description, a new class.

https://radar.anlyon.com/servers/io.github.acarchidi/agentforge/changes.xml

An Atom feed. Works in any feed reader or Slack's RSS app.

Badge

Anlyon Radar badge for io.github.acarchidi/agentforge
[![Anlyon Radar](https://radar.anlyon.com/badge/io.github.acarchidi/agentforge.svg)](https://radar.anlyon.com/servers/io.github.acarchidi/agentforge)

Source: MCP Registry · repository. Synced Nov 9, 2026, 21:17 UTC.

agentforge MCP server: questions and answers

Answered from the scan above, not from the publisher's description.

Is the agentforge MCP server safe?

No scan can promise that. What its definitions show: 29 tools, 22 of which can change, send or delete something, and 3 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.

Can the agentforge MCP server delete data?

No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.

Can the agentforge MCP server do things that cannot be undone?

Yes. 2 tools can send, pay, publish or merge: preflight and send_feedback.

What tools does the agentforge MCP server have?

29 tools: 2 irreversible, 20 write, 7 read. preflight, send_feedback, token_intel, token_research, contract_docs, contract_monitor, tx_decode, approval_scan and 21 more.

Does the agentforge MCP server need sign-in or OAuth?

Its remote endpoint listed its tools to an anonymous client, without sign-in.

Does the agentforge MCP server send data to outside services?

2 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): preflight and send_feedback.

Has the agentforge MCP server changed its tools?

Yes, 5 times since Anlyon Radar started tracking it, most recently on Nov 4, 2026. Every change is on this page and in its Atom feed.