Founder posts · SalesBuilder

What you get

3 posts · Each week, three posts in your voice for LinkedIn and X, written from what you shipped and learned, every number from a record. · Post
posts[] · Each post: the account, when it goes up, the text, and the record it came from.
held[] · posts that failed the grade twice, with the finding.
skipped · Topics left out, such as a change with no user-facing effect.

Connect once: LinkedIn, X account. How long: Minutes; posts go up through the week.

Trigger

It runs weekly at 08:00 on the first weekday.

TypeScript

every('mon 08:00', ($) => $.Post.write({ handles: ['linkedin:founder', 'x:founder'], perWeek: 3 }))

Steps

Your agent does these steps, one row per step; each step’s check is part of its row and its price.

Step	Verb	Returns	Price	Approval
1. Read the week’s material	`Collection.query`	Call notes, shipped changes, answers that worked and the latest test readout	Free	Runs on its own
2. Choose three topics	`Content.draft`	Three topics a buyer would stop for, each tied to one record	About 0.2¢	Runs on its own
3. Write each LinkedIn post	`Content.draft`	A post per topic, in your voice, with the record’s numbers	About 3¢	Runs on its own
4. Cut each into an X version	`Content.draft`	The same point in X’s length, with the link in a reply	About 3¢	Runs on its own
5. Schedule them on your accounts	`Post.schedule`	Each post scheduled at the account’s best slot	Free	Runs on its own

Calls: Collection.query, then Content.draft, then Post.schedule.

About 6.2¢ per 3 posts, each time it runs, capped at 25¢.

Example: give it a goal

Write three LinkedIn posts this week from what I shipped, spend up to 50¢, and show me the plan first.

The plan it shows runs 5 steps, priced before it spends anything; Run it above shows the result it returns.

When something fails

Each failure, and what happens next

When	Then
The week has no new material	Nothing is scheduled and the readout says so; it never invents a topic.
A post fails the grade	It is rewritten once; a second failure holds it with the finding, and it is not billed.
An account needs signing in again	Scheduling stops on that account with the fix line; the others go ahead.
You call it twice in one week	The week’s posts are returned: it is idempotent per handle, ISO week and slot.

Each step retries three times, then falls through to the next route; one item's failure lands in skipped with its reason, never failing the call. Before a step would cross maxCost it stops and returns what it has, with a resume link. Each step's idempotency key is the call's key plus the step and the item: leaseline-post-write-1-<step>-<item>.

How it is graded

A second model reads each post beside your last 20 posts and the record it cites. It passes when it reads as you on the voice samples, every number matches its record, and it makes one point. A miss is redone at no charge.

Change it

Ask	Compiles to
Only LinkedIn, and only from what I shipped.	`Post.write`
Harvest the people who comment, two days after each post.	`Audience.harvest`
Turn the best post into a page on my site.	`Page.answer`

TypeScript

import { $, workflow } from 'startup-builder'

export default workflow({
  name: 'post-write',
  verb: 'Post.write',
  input: { handles: 'string[]', perWeek: 'integer', from: "('calls' | 'changes' | 'answers' | 'experiments')[]" },
  trigger: { schedule: 'mon 08:00' },
  maxCost: 0.25,
  grade: { evaluator: 'your last 20 posts; the cited record', passWhen: 'reads as you; every number matches; one point', remedy: 'redo' },
  execute: async (ctx, { handles = ['linkedin:founder', 'x:founder'], perWeek = 3, from = ['calls', 'changes', 'answers', 'experiments'] }) => {
    const week = await ctx.step('source', () => $.Collection.query({ from, since: '7d' }))
    if (!week.rows.length) return { posts: [], skipped: 'nothing new this week' }
    const posts = await ctx.step('write', () => $.Content.draft({ kind: 'founder-posts', from: week.rows, count: perWeek, handles, maxCost: 0.1 }))
    return ctx.step('schedule', () => $.Post.schedule({ posts: posts.map((p) => p.$id), slots: 'best' }))   // graded before it schedules
  },
})

Extend it with extends: 'Post.write' and base(), or compose your own from these steps.

Works with

Plays that order it: Reuse what you already published, Build in public, Category kit, Founder voice, Organic social, Owned newsletter.

It serves Publish and post. It deepens Content marketing.

For agents

For agents: the guide, its approvals, retries and lookups

Fact	Value
Guide	`Post.write`
Guide approval	`limits`
Goal	Write three LinkedIn posts this week from what I shipped, spend up to 50¢, and show me the plan first. It may change counts, the order of steps, soft filters and the route a read takes, and it stops early when the goal is met. It never relaxes your words, raises its own budget, answers its own approval, calls a Play or acts outside your standing approval.
Trigger	`every('mon 08:00', ($) => $.Post.write({ handles: ['linkedin:founder', 'x:founder'], perWeek: 3 }))`
Retry	Three times with backoff, then the next route; the same idempotency key never charges, sends or posts twice
Idempotency key	`leaseline-post-write-1`
Lookups	A `find:` input asks unless exactly one record fits

Step
1. Read the week’s material
2. Choose three topics
3. Write each LinkedIn post
4. Cut each into an X version
5. Schedule them on your accounts

The amounts live once, on the hub (approvals).

Who does each step, and how it is checked

Step	Function	Doer	Checker	Assurance	Inputs
1. Read the week’s material	`code`	Code	Code checks each result against its declared type	`measured`	`$0.from`
2. Choose three topics	`agentic`	The Content lead agent	A second model grades the output	`graded`	`$1`
3. Write each LinkedIn post	`agentic`	The Content lead agent	A second model grades the output	`graded`	`$2.topics, $0.voice`
4. Cut each into an X version	`agentic`	The Content lead agent	Adversarial review by a second model: Pass, or the claim cut	`graded`	`$3.posts`
5. Schedule them on your accounts	`code`	Code	Code reads the result back from where it landed	`measured`	`$4.passed, $0.handles`

The doer is never the checker, and the weakest assurance on the path is the whole call’s.

Under the hood

Under the hood: the Systems, Nouns and Events it touches

Kind	What it touches
Systems	Sites and content, Social
Nouns	Collection, Content, Action, Post
Events	`Collection.queried`, `Content.drafted`, `Action.evaluated`, `Post.scheduled`

Reference

Inputs, outputs, its Event and its errors

Name	Type	Default	Meaning
`handles`	`string[]`	`['linkedin:founder', 'x:founder']`	The accounts it posts as.
`perWeek`	`integer`	`3`	Posts a week, at most. Fewer when the week had less to say.
`from`	`('calls' \| 'changes' \| 'answers' \| 'experiments')[]`	`['calls', 'changes', 'answers', 'experiments']`	Where topics come from: call notes, shipped changes, answers that worked, test readouts.
`voice`	`File`	`your last 20 posts`	Samples of how you write; the voice grade reads against them.
`maxCost`	`number`	`0.25`	The most this call spends each time it runs, in USD. It stops before a step would cross it and returns what it has, with a `resume` link.

Emits Post.written. Refuses with max-cost before it spends, and with the Data calls and Accounts groups (refusals).

What you getCopy Anchor Link

TriggerCopy Anchor Link

StepsCopy Anchor Link

Example: give it a goalCopy Anchor Link

When something failsCopy Anchor Link

How it is gradedCopy Anchor Link

Change itCopy Anchor Link

Works withCopy Anchor Link

For agentsCopy Anchor Link

Under the hoodCopy Anchor Link

ReferenceCopy Anchor Link

Contents