agentleFS
matthewdonsemail-labmatthewdonsemail-lab/log/.claude/skills/log-twitter-manager-maxxing/SKILL.md

log-twitter-manager-maxxing

Skill63 starsChanged yesterdayMITView on GitHub

Manager maxxing: the managers of the biggest Kick streamers. A log play (docs/useCases/twitter/managerMaxxing.md): The owner's go-to-market, in his words (https://x.com/matthewsoldit/status/2118285429674017268): \"find every single kick streamers managers on twitter, reply guy them for a few days then cold dm them to use log and show claviculars twitter\". The pitch, once they're talking: if Claude Code explains the sponsor deals a streamer does, with tracked links, sign-up rates go up, which means more affiliate money. Letting AI use your social media increases your surface area of luck.

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 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.

  • 0

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Every send goes through the account's one queue, gated against his voice, and a refusal from X stops the run (rules 94, 105).

    line 87

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 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

Manager maxxing: the managers of the biggest Kick streamers

This is the log play twitter/managerMaxxing. Before anything else, call the log MCP tool getGuide({ useCase: "twitter/managerMaxxing" }): 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).

The owner's go-to-market, in his words (https://x.com/matthewsoldit/status/2118285429674017268): "find every single kick streamers managers on twitter, reply guy them for a few days then cold dm them to use log and show claviculars twitter". The pitch, once they're talking: if Claude Code explains the sponsor deals a streamer does, with tracked links, sign-up rates go up, which means more affiliate money. Letting AI use your social media increases your surface area of luck.

The people this is for (the ICP) are the managers of Kick streamers, category by category, starting with the streamers with the most viewers. It runs on the network-maxing logic (networkMaxing.md): every touch is about them and what they're doing, and the ask comes only once they're talking.

How the owner asks for it: "manager maxxing", "find every kick streamer's manager on twitter", "reply guy the kick managers with memes", "the managers that engaged, send them their tweet, a message and a voice note".

Run it from a fresh context

Say: "go on docs/useCases/twitter/managerMaxxing.md for <category, or all of them>". The agent reads this page, README.md and networkMaxing.md, loads getGuide({ topic: "xSending" }) and getGuide({ topic: "extensions" }), and runs the steps.

The steps

Find the agencies, their talent and their managers

The way in is the agency. Category by category, the top streamers lead to the agencies that manage them, each agency leads to the rest of its talent, and its staff are the managers. Agencies sit next to other agencies (talent they follow is managed by another), so each one found opens the next.

#StepToolDone when
1The top Kick streamers by viewers, grouped by category. StreamerStats ranks the last 30 days at streamerstats.com/kick/channels?period=30d (pages 1 to 3, the top 75, are open without a login), each row with its category, and each channel's page at streamerstats.com/kick/channels/<name> links the streamer's X. A category's own page only shows who is live now. Streams Charts is behind Cloudflare and KickStats renders with JavaScriptthe built-in browser on StreamerStats (rule 53: the source is named)~/.logCli/managerMaxxing/streamers.json: rank, category, viewers and X handle per streamer
2Each streamer's agency from their X bio: a business email's domain (@<agency>.gg, @<agency>.com), an agency's handle (@<agency>), or a named manager ("Contacto y Publicidad @<manager>")treg treg.x.user.profileeach streamer's agency, and how it was found
3Each agency's roster from its own site (the creators it lists), and its own X accountthe built-in browser on the agency's site, treg anyapi.twitter.search_users { query: "<agency name>" }the agency's talent and its X handle
4The agency's people: who the agency's X account follows, kept where the bio names a role there ("Senior Talent Manager @<agency>", "Talent Manager @<agency>", "Agent @<agency>"). The same list shows more of its talent (their business emails on its domain) and the agencies next to it (talent it follows managed by another domain), which go back into step 3treg tikhub.x.twitter-web-fetch-user-followings { screen_name }~/.logCli/managerMaxxing/agencies.json: per agency its talent, its people with their role, and the adjacent agencies

Reply guy them for a few days, with memes

<!-- stepsDiagram: written by `bun run docs:build` from the steps table below, not edited by hand -->
flowchart TD
  s0["5. Each manager's recent original posts<br/><i>treg treg.x.user.posts username</i>"]
  s1["6. A reply in the owner's voice about their post, with a meme pick…<br/><i>memelordSearch, voice</i>"]
  s2["7. Into the account's one queue as replies, spread over the days,…<br/><i>twitterQueueBuild, twitterQueueCheck, twitterQueueApprove, twitterQueuePlan, twitterQueueRunLive</i>"]
  s0 --> s1
  s1 --> s2
<!-- /stepsDiagram -->
#StepToolDone when
5Each manager's recent original poststreg treg.x.user.posts { username }a post per manager worth replying to
6A reply in the owner's voice about their post, with a meme picked to his taste for that moment, uploaded as media. Gate every replymemelordSearch, voice { to, context, text }every reply passes
7Into the account's one queue as replies, spread over the days, a few minutes apart (rule 105). Each reply likes the post and follows them firsttwitterQueueBuild { action: "reply" }, twitterQueueCheck, twitterQueueApprove, twitterQueuePlan, twitterQueueRunLiveX's answer is sent

Who engaged, then the DM

<!-- stepsDiagram: written by `bun run docs:build` from the steps table below, not edited by hand -->
flowchart TD
  s0["8. Who engaged back: liked or replied to a reply, followed the own…<br/><i>twitterNotifications, twitterFollows, twitterDms</i>"]
  s1["9. To each one who engaged, the cold DM the network-maxing way: on…<br/><i>voice, twitterQueueBuild</i>"]
  s2["10. Then a brief voice note in the owner's cloned voice, once the o…<br/><i>voice, twitterQueueAdd</i>"]
  s3["11. When they answer, the opening message: get to know them and the…<br/><i>twitterDms, voice, twitterDmLive</i>"]
  s0 --> s1
  s1 --> s2
  s2 --> s3
<!-- /stepsDiagram -->
#StepToolDone when
8Who engaged back: liked or replied to a reply, followed the owner, or messaged himtwitterNotifications { account }, twitterFollows, twitterDms { account, with }the managers who engaged
9To each one who engaged, the cold DM the network-maxing way: one of their own tweets forwarded first, then one line about it. The opener only, never the askvoice { to, context, text }, then twitterQueueBuild { action: "dm" }every opener passes
10Then a brief voice note in the owner's cloned voice, once the openers have gone: one spoken question that goes a step further on their tweetvoice { to, context, text, spoken: true }, then twitterQueueAdd { action: "voice" }X's answer is sent for each voice note
11When they answer, the opening message: get to know them and the streamers they run, then, in the conversation, log: Clavicular's X as the example, Claude Code explaining sponsor deals with tracked links, more sign-ups, more affiliate moneytwitterDms { account, with }, voice, twitterDmLivethe conversation is going

What to know

  • The ICP is the manager, not the streamer. A streamer only matters here as the way to their manager.
  • Never guess a manager. Only a bio that names the role at the agency counts. Searching X posts for "<streamer> manager" or "<agency> talent manager" brings up jokes, fans and recruiters (on 2026-11-05 both did), and X's people search matches names, not bios; the agency's own following list is what finds its staff.
  • Reply guy first, for a few days, then the DM only to the ones who engaged. Someone who never engaged isn't DMed yet.
  • The opener is never the ask. log, Clavicular's X and the affiliate money come up once they're talking.
  • Replies go to authors with 1500 followers or fewer; a bigger manager needs the owner to name them.
  • Every send goes through the account's one queue, gated against his voice, and a refusal from X stops the run (rules 94, 105).

Next

  • Steps 1 to 4 are read by hand from StreamerStats, the agencies' sites and treg for now; a command that builds streamers.json and agencies.json would make them repeatable.
  • StreamerStats shows only the top 75 without a login; a deeper ranking per category needs the owner signed in there.

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.