LLM Red-Team Scanner

Deterministic LLM red-team scan: prompt injection, jailbreak, system-prompt extraction and MCP tool abuse in agent prompts.

Transfer logs for one token contract, optional holder filter, amounts decoded with token decimals, read-only.

Contract or EOA by address: code size, token name/symbol/decimals/supply, EIP-7702 delegation, read-only.

Transaction by hash: from, to, value, nonce, gas used, fee, status and calldata selector, read-only.

Raw eth_getLogs output filtered by contract, topic and block range, read-only.

Extract the readable article content from any url as clean markdown with a sha256.

Deterministic LLM red-team scan: prompt injection, jailbreak, system-prompt extraction and MCP tool abuse in agent prompts.

Jailbreak, guardrail and MCP tool-abuse scan: 0-100 risk score with printable evidence.

Deterministic LLM red-team scan: prompt injection, jailbreak, system-prompt extraction and MCP tool abuse in agent prompts.

Monitor a page for changes: url change monitor with before/after sha256 and the diff.

ERC721 or ERC1155 standard, collection name, symbol and tokenURI by contract and token id, read-only.

Current ERC721 owner by contract and token id via ownerOf, read-only.

Did this page change, plus before and after sha256 and the added/removed lines.

Convert any public page to clean markdown plus the sha256 of the extracted text, so an agent can read a link without a browser.

Cheapest paid request on this seller: proves your 402 challenge, payment signing and settlement work. No upstream dependency, so a failure means your client, not the seller.

Same asset priced across exchanges and aggregators in one call: per-source deviation from the median in bp, spread, TIGHT or DRIFT verdict.

44-probe deterministic LLM red-team scan: 20 attack classes, 0-100 risk score, CI gate.

Same 44-probe engine on the security-audit route: 0-100 risk, printable evidence, CI gate.

One-probe penny test of the full red-team kit: 0-100 risk score, instant delivery.

Deterministic LLM red-team scan: prompt injection, jailbreak, system-prompt extraction and MCP tool abuse in agent prompts.

Fetch any page as clean markdown plus the sha256 of the extracted text.

Neural-voice text to speech: pass a string, get an mp3 plus sha256. No API key, no signup.

Receipt by hash: status, gas used, fee paid and emitted log count, read-only.

Did this page change, plus before and after sha256.

Current weather plus 24-hour forecast for a coordinate: temperature, humidity, precipitation, wind speed. No API key, no signup.

Your rank and percentile by settled calls across the ranked market, plus your per-buyer economics against the market medians.