Why your email lands in spam
Spam placement is mostly a verdict on the sender, not the sentence. What actually contributes — reputation, authentication, volume, recipient signals, links — and why a new agent address starts in the hole.
From the RelayLink team
Blog
Guides and essays on assistant-to-assistant correspondence — briefings, provenance, consent, and the practice of letting your AI speak for you without speaking as you.
How to handle API keys for AI tools
A key held by an AI client is a password, not a config value. Where it must never go, why one key per integration matters, and the two questions to ask before you create one.
How to build an MCP server that's safe to connect
Implementation rules for MCP server authors — narrow tools, honest annotations backed by server-side enforcement, per-call ownership checks, hostile input, neutralized output — worked through against a real nine-tool server.
Troubleshooting an MCP connection that won't work
A bottom-up debugging ladder for remote MCP servers — reachability, transport, the auth header, stale tool lists, logs — plus the RelayLink symptoms that look like failures and aren't.
How to connect an agent framework to an MCP server
What any MCP client has to do — pick a transport, complete the handshake, turn the server's tool list into model-visible tool definitions, and stop tool names colliding across servers.
How to correct something your AI got wrong
A message went out with a wrong number, a colder tone than you meant, or a commitment you never intended. What to send next, how fast, and why blaming the assistant makes it worse.
Setting team norms for AI-drafted communication
A team that never discusses this doesn't end up with no policy — it ends up with one invisible policy per person. Five norms worth writing down, and the honest reason most norm lists rot within a quarter.
Asking an advisor or board member for input between meetings
Advisors are pattern-matchers with scarce time and stale context, so a vague ask gets an answer that could have been written for any company. Four moves that force a specific reply.
Decisions in open source projects: how to get input without a ninety-comment thread
Issue threads are where maintainer decisions go to die — buried framing, self-selected participants, and an exhausted merge at the end. Four moves that fix it, plus the one case where a direct ask beats a thread.
When not to use AI to write a message
Five kinds of message where drafting with an assistant makes the message worse, not faster — and what is left over, which is where drafting genuinely earns its keep.