Sort your unread messages on LinkedIn, using one prompt

What do I actually type?

This one sentence. That is the whole thing.

How does it work, step by step?

Paste these one at a time. Each one is written in normal English — no command names, no flags.

We gave these exact prompts to a real AI assistant and wrote down every command it ran. It chose 2 commands: list-conversations, get-conversation-messages. Nobody told it which ones to use.

What it printed when we ran it

Not a mock-up. This is the actual run: the assistant was given the prompts above and nothing else, and these are the 2 distinct commands it chose from the catalogue. It ran read-only on a real LinkedIn account — nothing was sent, posted or invited.

Can I make it run on its own?

Yes. Add one more sentence and it keeps happening without you. This is the part a chat window cannot do — a saved job runs on the days you are busy.

How many LinkedIn actions does this use?

LinkedIn limits how much any account can do in a day. These are the real numbers, and your assistant checks what is left before it starts.

• Inbox reads: uncapped. The one budget category the platform leaves open. Healthy accounts sustain roughly 240 reads a minute and remain healthy. That is what makes a nightly full-inbox pass sensible rather than reckless — the read side has no ceiling to size against.
• Messages: 150 a day, reset midnight UTC. Reading the inbox is free. Sending is not. Two hundred threads in, roughly ten per cent worth a human reply on the usual ratio, is twenty replies against a budget of a hundred and fifty — about a seventh of the day. The cap is not what stops you. Judgement is.
• Writes: 15 a minute. The velocity ceiling that sits under the daily total. Twelve replies leaving in one burst is inside every daily cap and still reads as a mail-merge to twelve people who each know they wrote something different.
• Profile views: 120 a day. Untouched by the triage itself — a conversation read is not a profile read — but worth naming because it is the budget that shapes almost every other job on the account. If you enrich the sender before drafting, that is where the cost lives.

Watch it instead

5 minutes

Common questions

What if the classifier lumped everything into "buyer interest"?

It classified on job title rather than content, so every VP landed in the same bucket regardless of what they wrote. Restate the archetypes in terms of what the person is asking for, not who they are.

What if a draft answers a question they did not ask?

It worked from the last message alone rather than the full thread. The skill file tells the assistant to open the full thread before drafting on anything hot. if a draft misses context, the sync ran on a stale copy or the fetch step got skipped.

What if the morning file is empty?

The nightly job did not run, or the LinkedIn session needs reconnecting and failed quietly. Ask the assistant to check the account status before assuming the tool broke.

What if it acted as the wrong account?

Almost always a stale saved setting beating the one you thought was in effect. Ask which account it is using; it can report where that answer came from, and rehearse any send before it goes out.

How many LinkedIn actions does sort your unread messages use up?

Inbox reads: uncapped. The one budget category the platform leaves open. Healthy accounts sustain roughly 240 reads a minute and remain healthy. That is what makes a nightly full-inbox pass sensible rather than reckless — the read side has no ceiling to size against.

How do I set this up?

Once, and it takes about four minutes. After this you only write sentences.

$ npm install -g @connectsafely/cli$ export CONNECTSAFELY_API_KEY=...$ csly list-linkedin-accounts$ export CONNECTSAFELY_ACCOUNT_ID=...$ csly skill install

The package is @connectsafely/cli on npm. You need Node 20 or newer.

Then start your chat with this line, so your assistant uses the CLI instead of trying to open LinkedIn itself and hitting a login page: