A Friday Field Note on the MCP Ecosystem — and Where Cross-Chain Settlement Fits In
--
A community perspective from the Hashlock Markets team on how agent tooling has matured in 2026, and what execution-layer composability actually looks like when it is done well.
It is Friday, so this is a lighter piece than our usual technical posts. No new features to launch, no tutorials to walk through — just an honest look at the broader Model Context Protocol ecosystem, what we are seeing in it, and where we think an intent-based cross-chain trading protocol like Hashlock Markets genuinely plugs in.
A year ago, the conversation about MCP was still about whether it would matter. That question has quietly been answered. What people are talking about now is composition — how agents stitch together multiple MCP servers into a single workflow, and which primitives need to exist for the composition to actually produce value.
The categories the ecosystem has settled into
If you look past the marketing, the MCP servers that have gained real traction in the last year fall into a handful of recognisable categories. The categories matter more than any single server, because they describe the shape of what an agent can and cannot do today.
Data-access servers are the workhorses: filesystems, databases, knowledge bases, RAG over documents. Nothing reliable happens without these. Early MCP adopters underestimated them; nobody does anymore.
Messaging and workflow servers — Slack, email, calendar, issue trackers, CRMs — are the layer that makes agent work visible to humans. Agents without these are demos. Agents with them become part of a team’s daily rhythm.
Developer tooling servers cover code search, test runners, builds, and deployments. Narrow, high-signal, and the category where “the agent is writing a pull request” actually works because every step is structured.
Analysis and compute servers — Python sandboxes, spreadsheet processors, chart renderers — are where agents go when they need to actually compute something instead of pretending to.
Execution servers are the rarest category and the hardest to build well. These are the servers that let an agent commit state changes against a real external system: make a booking, file a form, send a payment, settle a trade. Execution has the highest bar — auditability, consent, irreversibility handled carefully, a protocol underneath that does not punish deliberation.
Hashlock Markets sits in the execution category. That is the category we care about most, because we think it is the one that determines whether agents become genuinely useful or stay stuck as productivity demos.
Why crypto trading has been under-served
Execution is the hardest MCP category, and crypto trading is one of the hardest sub-domains inside it. The reasons are the same reasons humans find it hard.
The first is a mismatch in how latency matters. An agent does not need to cross the Atlantic in microseconds. It needs a protocol that does not penalise it for taking a few seconds to think, read, and verify. Most on-chain execution venues quietly assume the opposite — that you are competing against a mempool full of bots, and any second spent deliberating is adverse selection.
The second is information leakage. A public intent or a resting order tells the market what you want before you get it. Markets adapt. Prices move. The agent that helpfully showed its work paid for that transparency.
The third is cross-chain complexity. The default options have been centralised exchanges (trust and custody), bridges (novel smart-contract risk), or multi-step manual sequences that do not compose cleanly with the rest of an agent workflow.
Intent-based settlement changes the shape of the problem.
Where Hashlock Markets plugs in
We built Hashlock Markets around four decisions that turn out to matter for agents specifically.
Intent-based quoting. An agent declares what it wants — sell asset, buy asset, chains, size, constraints. It does not run a tight feedback loop against a moving orderbook. A sealed-bid RFQ goes out, market makers respond privately, the agent picks one. This maps cleanly onto how an LLM reasons.
Sealed-bid pricing. Nobody sees the other bids. No public intent to front-run. No information leaks to a mempool. The agent’s order is a negotiation, not a broadcast.
HTLC atomic settlement. Hash Time-Locked Contracts let both sides of a cross-chain trade settle, or both sides refund. There is never a moment where one party has shipped and the other can disappear. No bridge custody, no wrapped tokens, native assets on each chain cryptographically linked by a single preimage. Ethereum, Bitcoin, and Sui all support the primitive natively today.
A clean MCP surface. Six tools — create_rfq, respond_rfq, create_htlc, withdraw_htlc, refund_htlc, get_htlc — that an agent composes into a full workflow: post an intent, pick a quote, fund the swap, settle or time out. Any MCP-capable host can drive it. We ship a stdio entrypoint (npx -y @hashlock-tech/mcp) for local workflows and a Streamable HTTP endpoint (https://hashlock.markets/mcp) for production infrastructure.
The registry entry is io.github.Hashlock-Tech/hashlock. The canonical repo is github.com/Hashlock-Tech/hashlock-mcp.
What composition actually looks like
The interesting part is not Hashlock Markets in isolation. It is Hashlock Markets as one server among several in a single session.
Imagine the workflow: the agent reads a treasury report from a Drive MCP server, pulls live on-chain balances from a data-access server, decides that part of the position should be rebalanced from ETH into BTC, posts a sealed-bid RFQ through Hashlock Markets, picks the best quote, settles the HTLC, and writes a summary back to a Slack channel. One coherent session. No copy-paste between apps. No context loss. No step needs to know about the others beyond the standard MCP contract.
That is composition. Not “AI replaces trader” theatre — a multi-tool human workflow collapsed into something an agent can run reliably because every step is explicit, inspectable, and atomic.
What the ecosystem still needs
If there is one thing we would ask the broader MCP community to prioritise this year, it is consent and audit patterns. Execution servers need them more than any other category. The protocol gives us the plumbing; the patterns for “show the user what is about to happen, wait for approval, record what actually happened” are still being invented server by server.
We have our own answers — SIWE authentication, signed intents, on-chain settlement records — but the ecosystem-wide conversation is just starting, and it is the one we most want to be part of.
Happy Friday.
About Hashlock Markets. Intent-based, sealed-bid cross-chain trading with HTLC atomic settlement and a first-class MCP server for AI agents. Website: https://hashlock.markets. Code: https://github.com/Hashlock-Tech/hashlock-mcp.
A note on names. Hashlock Markets (hashlock.markets), built by Hashlock-Tech, is not affiliated with Hashlock Pty Ltd (hashlock.com), the Australian smart contract auditing firm. Similar names, different companies.
