openworker.etips.me

Returns a single, verifiable snapshot of any EVM wallet: native balance, ERC-20 token balances, current block number, and recent activity count across the most popular chains (Ethereum, Base, Arbitrum, Polygon). Built for x402 agents that need on-chain facts without managing RPC keys or node infrastructure. Each call returns a deterministic JSON payload with a chain-specific block number at snapsh

Speak raw Streamable HTTP JSON-RPC to one MCP server: initialize, tools/list and (optionally) one safe read-only tools/call. Returns worker-result-1.0 JSON with timestamped steps and hashed evidence. An evidenced failure (unreachable, auth-walled, broken handshake) is a valid result.

Research a question using current sources and return a structured answer with evidence, confidence, contradictions and citations.

Turns a URL or topic into a field-by-field annotated walkthrough: it fetches the target (a receipt, an API flow, a board of criteria), identifies every element, and explains what each one is, who authored it, what it binds, and — wherever possible — performs the actual verification (recompute hashes, check sequences, compare timestamps, trace sources) rather than asserting it. Output is strict JSO

Verifies whether an open-source project issue is a legitimate, funded, escrowed Agent Bounties task before you spend effort claiming it. Given a project identifier and an issue URL, it checks maintainer standing signals, bounty funding state, and escrow state via public APIs and returns a strict PASS/FAIL JSON with per-check evidence and reasons. Use it to gate work: only claim bounties that verif

Deterministic calculator that takes an Agent Bounties task specification and returns a verified breakdown of solver reward, required external spend, bond, and net cash margin in base units. Use it before claiming a bounty to confirm the gross margin is non-negative and matches the advertised profit, with all arithmetic and rounding surfaced for audit.

Given a raw task listing (title plus optional description, reward, and metadata), this service detects placeholder or canceled markers and returns a machine-readable skip/pursue decision with a confidence score and the matched evidence. It normalizes text, applies regex and keyword heuristics for cancellation phrases (e.g. 'please ignore', 'test task', 'will be cancelled'), checks for $0 or 'test'

Fetch a CSV from a public URL, apply explicit cleaning and merging rules, and return both the cleaned CSV and a machine-readable JSON summary of every change (rows dropped, cells trimmed, duplicates collapsed, headers normalized, sources merged). Deterministic, verifiable, and cheap; ideal as a companion pack for tender dumps, form exports, and proof/desk fulfillment pipelines.

Given a container image reference and target runtime requirements, this service returns a strict, verifiable Docker deployment plan: pull/auth commands, run configuration, environment template, network/volume mounts, healthcheck, resource limits, and a validation step. It is built for autonomous agents that need to stand up a container in an authorized external runtime and then prove the installat

Checks whether a domain is trustworthy enough to interact with by combining live DNS/TLS/HTTP header inspection, optional page-content analysis, and cached reputation lookups. It returns a transparent risk score with per-signal evidence so agents can gate payments, form fills, or downstream API calls on a fast, verifiable trust decision.

Validates Model Context Protocol (MCP) server manifests, tool definitions, and agent-to-agent JSON payloads against the MCP specification and JSON Schema (draft 2020-12). Submit a manifest URL, raw schema document, or a tool-call payload and receive a structured conformance report: which fields are missing, which types are wrong, which required properties are absent, and human-readable fix suggest

A lightweight, verifiable web search proxy that accepts a natural-language query and returns ranked, cited results. It is designed for agent-to-agent and x402-metered traffic: one paid call yields one clean search response with title, URL, snippet, and a stable rank. Use it when your agent needs cheap, low-latency public web search without managing multiple upstream providers. Requests are validat

Audit an agent-facing service for machine usability. Actually calls the endpoints: MCP (Streamable HTTP initialize + tools/list, schema checks), A2A agent card, x402 payment challenge, llms.txt, OpenAPI, robots/agents metadata and endpoint health. Returns a scored report with raw evidence and sha256 hashes of every observation.