MCP security report
rpc.aex402.com
F13/100
Security grade FHigh-severity signals — review these before connecting.
Connected ✓33 tools scanned
https://rpc.aex402.com/mcp
highlaunch_token
Custodial signing and submission of on-chain financial transactions without user-held keys
The default (non-build) path has the aeX402 backend sign and submit a blockchain transaction on the user's behalf ('the aeX402 backend SIGNS + SUBMITS the createVPool tx for you'). This is a significant capability for a tool whose stated purpose is simply launching a token: it moves real funds/gas and creates an on-chain asset tied to an optional owner address without the calling agent ever obtaining a signature from the user's own wallet. This custodial-signing pattern is a much broader and riskier capability than the rest of the toolset (which generally returns unsigned instructions for client-side signing), and could be invoked by an agent without clear user consent to an irreversible on-chain action.
mediumpayment_help
References secrets or data exfiltration
The tool text references credentials, environment variables, or sending data elsewhere. This can be legitimate, but review what this tool actually accesses.
mediumagent_delegate
Spawns sub-agents with full tool access and recursive delegation
This tool creates independent agent sessions that have 'full tool access' and can recurse up to depth 3. This is a broad capability that could be used to fan out actions (including sensitive on-chain transactions) through child agents outside of the direct oversight/context of the primary conversation, making it harder for a user to audit what actions were actually taken and by which agent.
mediumsolana_rpc
Unrestricted passthrough to arbitrary RPC methods including state-changing calls
Allows calling 'ANY' Solana-compatible RPC method by name, with no allowlist/denylist distinguishing read-only queries from state-changing or sensitive methods (e.g., simulateTransaction, getProgramAccounts at scale). Combined with evm_rpc's equivalent 'ANY EVM JSON-RPC method' capability (including eth_sendRawTransaction), this gives much broader raw access to multiple chains than the purpose-built typed tools, increasing the blast radius if misused or driven by a manipulated prompt.
mediumevm_rpc
Unrestricted passthrough to arbitrary EVM RPC methods across many chains
Permits any EVM JSON-RPC method (including eth_sendRawTransaction) across 9 chains with no method allowlist. While submitting a transaction still requires an already-signed payload, the tool's unrestricted method surface is unusually broad relative to the typed, purpose-specific tools elsewhere in this set and could be used to bypass safer, narrower tool paths.
lowmcp_linq
Unusually long tool description
Very long descriptions are sometimes used to bury instructions to the AI. Skim it before allowing.
low
Not on gate's verified list
This server isn't on gate's handpicked list. That's not necessarily bad, but there's no third-party signal about it — review its tools carefully.
lowlaunch_token
Description explicitly references 'claude.ai user'
The description singles out 'a claude.ai user with NO WALLET' as the beneficiary of the custodial flow. Explicitly naming a specific AI assistant/platform inside a tool description is unusual and worth noting as a possible sign the description was crafted to specifically steer that assistant's behavior (here, toward using the custodial signing path) rather than being a neutral, audience-agnostic tool description.
lowapi_get
Loosely-validated free-form path parameter to backend API
The 'path' argument accepts an arbitrary string appended as a path+querystring to rpc.aex402.com, with the allowed endpoints only documented in prose (not enforced by the schema). This could allow probing of undocumented or unintended endpoints on that host beyond the listed read-only ones.
Embed this badge
Show your MCP server’s security grade
Paste this into your README:
[](https://gateturbo.com/report/cf8a19b7-5829-495f-9966-48478b9ffd84)Scanned 10/3/2026 · This is a point-in-time snapshot of the server’s public tools. Re-scan.
Connect this server safely with gate.
Continuous re-scans, drift alerts, per-tool allow/ask/block, and a log of every call.
Monitor it — free