agentleFS
matthewdonsemail-labmatthewdonsemail-lab/log/.claude/skills/log-twitter-follow-ups/SKILL.md

log-twitter-follow-ups

Skill63 starsChanged 2 days agoMITView on GitHub

Following up in your own voice. A log play (docs/useCases/twitter/followUps.md): Answer or follow up the people the owner picked, each in the owner's own voice, the way they write to that person. It's for warming a relationship as much as for a deal: say it was good to meet, ask for advice, take someone up on an offer, keep a door open. Nothing goes out until the owner has seen every draft and said yes.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Never ask something the person already told the owner.

    line 38
  2. Distinctive:

    Use only facts the owner said.

    line 39
  3. Very specific:

    Every call goes through the account's one X session (rule 105).

    line 43

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Following up in your own voice

This is the log play twitter/followUps. Before anything else, call the log MCP tool getGuide({ useCase: "twitter/followUps" }): it returns this page and makes it the session's play, so every log tool's answer says which step you are on and a send with no play loaded is refused (rules 157, 158). Then follow the steps in order.

Draft every message from that person's own thread and matt's real replies (voice with no text, search), never from memory, and never send two that read alike (sameAsSent).

Answer or follow up the people the owner picked, each in the owner's own voice, the way they write to that person. It's for warming a relationship as much as for a deal: say it was good to meet, ask for advice, take someone up on an offer, keep a door open. Nothing goes out until the owner has seen every draft and said yes.

How the owner asks for it: "follow up with these, not for anything in particular, just warming up the relationship", "say it was cool to meet him", "ask him for some advice on what he's doing", "take him up on the doc he offered".

Run it from a fresh context

Say: "go on docs/useCases/twitter/followUps.md", and who, with what the owner wants from each. The agent reads this page and README.md, loads getGuide({ topic: "xSending" }), and runs the steps.

The steps

#StepToolDone when
1Read each person in full, old DM thread and X Chat mergedtwitterDms { account, with: "<handle>" }everything they've said, oldest first
2Take what the owner wants from each, in the owner's words only (met in person, ask advice, take up an offer)the ownerone intent per person
3How the owner really writes to this person: their messages to them first, then the owner's whole historyvoice { to: "<handle>", with: "<handle>", context: "<what they said>" }the owner's own lines to draft from
4Draft each one from those lines, in the owner's words and length. Every draft reads differently: no shared opener, no stock line. Each line of a draft is one messagethe agentone draft per person
5Gate each draft against the same person, and rewrite until it passesvoice { to, with, context, text: "<draft>" }the gate passes
6Show the owner every draft as it will go out, one message per linethe ownerthe owner says yes, or what to change
7One person: dry run, then live. Two or more: put them in the account's one queue, never a loop of sends (rule 105): twitterQueueAdd { queue, targets, action: "dm", by } adds them, even while it runs, fitted between what is already planned; if no run is going, twitterQueueRunLive { queue, keepOpen: true, confirmLive: true } starts it. A draft on several lines goes out one message per linetwitterDm, then twitterDmLive { userId, text, approvedBy, confirmLive: true }, or twitterQueueAdd and twitterQueueRunLiveX's answer is sent for every part

What to know

  • Never ask something the person already told the owner. Read far enough back first (step 1, and search { query, with } for older talk).
  • Use only facts the owner said. Nothing invented about a meeting, a deal or a number.
  • --asOne (asOne: true) sends a draft as one message, only when the owner asks for that.
  • A drop or a 429 stops the rest: what went out stays out, nothing is resent, and the next send waits for the owner (rule 94).
  • An account has one queue and one run for everything it sends: follow-ups, reply-guy replies and answers go into the same one (rule 105). Never start a second queue or run beside it, and never send from a hand-run command or a loop (rules 105, 163). Each target carries its own text, and its dm items split one message per line the same way.
  • Every call goes through the account's one X session (rule 105); the answer's xSession names it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.