Skip to content

Autonomous development for real products. Governed agents build from specs you approve, follow the rules you set, and nothing merges until another agent proves it. Your decisions, judgement and taste are the bar that work must clear. Declare your vision and let agents build it.

docs(readme): a friendlier opening, the board's own badges [I10,I12]

Corey Olson9e59df4414bc
diff --git a/README.md b/README.md
index 4c1d243e..ad887da1 100644
--- a/README.md
+++ b/README.md
@@ -7,26 +7,34 @@
<h1 align="center">Pullboard</h1>
-<p align="center"><a href="https://pullboard.dev"><b>pullboard.dev</b></a> · lives in your git repo · no account · no dependencies · never calls a model</p>
+<p align="center"><a href="https://pullboard.dev"><b>pullboard.dev</b></a> · lives in your git repo · no dependencies · <b>Walk away, and it builds.</b></p>
<p align="center">
- <a href="https://github.com/pullboard-dev/pullboard/actions/workflows/gate.yml"><img alt="gate" src="https://img.shields.io/github/actions/workflow/status/pullboard-dev/pullboard/gate.yml?branch=main&label=gate&style=flat-square"></a>
+ <a href="https://pullboard.dev"><img alt="pullboard: 893 items" src="https://img.shields.io/badge/pullboard-893%20items-2f7d5b?style=flat-square"></a>
+ <a href="https://pullboard.dev"><img alt="verified: 100%" src="https://img.shields.io/badge/verified-100%25-6e7781?style=flat-square"></a>
<a href="https://www.npmjs.com/package/pullboard"><img alt="npm" src="https://img.shields.io/npm/v/pullboard?style=flat-square&color=121a17"></a>
- <img alt="node" src="https://img.shields.io/node/v/pullboard?style=flat-square&color=121a17">
<a href="LICENSE"><img alt="MIT license" src="https://img.shields.io/npm/l/pullboard?style=flat-square&color=6b4fc8"></a>
</p>
-**Accumulated judgement for agents.** Agents are good at tasks and bad at projects: every new session starts without what you decided, so you become the glue. The agents scale; your judgement doesn't. Pullboard is a structural approach to orchestrating agents coherently over long-term builds, so you can trust your agentic workforce to fill in the blanks the way you would, without re-prompting them.
+### Pullboard enforces your intent and decisions. Build autonomously.
+
+1. **Vision:** what you want.
+2. **Specification:** your vision written down.
+3. **Inspection:** independent verification.
+
+Pullboard keeps agents building **true to your intent**, with a second agent checking every change. Your judgement carries across every agent and every session, so a complex project stays coherent for months. It's a coordination queue for serious software: you decide what matters once, and many agents build, check and land the work.
<picture>
<source media="(prefers-color-scheme: dark)" srcset="docs/img/holds-dark.svg">
<img src="docs/img/holds.svg" alt="How it holds. Your decision becomes a bar each piece of work is frozen against; a second agent checks the exact commit; the board keeps what passed, what failed and why; the next agent starts from there.">
</picture>
-**Vibe code a real product.** Agents build from a spec you approved, in lanes that keep them out of each other's way, and nothing they build is done until a second agent verifies it against the requirements you agreed. Pullboard is a coordination queue for serious software. It's how we develop our quantitative systems.
+**Vibe code a real product.** Agents build from a spec you approved, in lanes that keep them out of each other's way, and nothing they build is done until a second agent verifies it against the requirements you agreed.
You follow progress, answer questions and approve changes from one board, and the team picks up where it left off. Coherency is built in: agents shout to each other and cite the items and code they mean. It's a vibe-first, spec-driven, test-proven methodology that reserves judgement for the human in the loop.
+https://github.com/user-attachments/assets/91bb722e-deb7-4b77-8512-5adc4fd22269
+
## Software factory
**Prompting is for discovery. Declaring is for delivery.** Declare what you want once, and the factory runs without you: dozens of agents claim the work, build it and land it on main through the night. Your job moves from typing prompts to the judgements only you can make. Your decisions, judgement and taste are the bar that work must clear. You come back to shipped features, not a chat log.
@@ -36,7 +44,7 @@ You follow progress, answer questions and approve changes from one board, and th
<img src="docs/img/factory.svg" alt="One night on Pullboard's own board, 16 October 2026: 38 agents, 9.6 hours from 11:17 pm to 8:51 am, with no moves from the person. 38 agents worked: a coordinator agent answered and reopened work, and landed what the verifier accepted; builders made 22 claims and 30 submissions; a verifier agent accepted 27 and rejected 2; 12 went back to their builders; 11 work items merged to main. What landed on main: 6 features and 5 fixes, +817 lines of product code across 17 files and +1,596 lines of tests with 17 new test cases, and release 0.8.27 on npm at 7:56 am. Among the features: The view signs in and links every board; pullboard.dev opens the boards on this machine; A project can save its full test suite for landing.">
</picture>
-### One agent. Or a hundred. (Maybe even thousands.)
+### One agent. Or a hundred. (Maybe thousands.)
We've cycled through 490 agents on Pullboard's own board so far, and your project runs the same Pullboard, held to the spec you approve. Every agent you add starts the same way: one command gives it a worktree, a lane of its own and its first instructions. Pair your phone when you're ready to walk away.
diff --git a/test/ci.test.js b/test/ci.test.js
index cba15b07..2f8dcdfc 100644
--- a/test/ci.test.js
+++ b/test/ci.test.js
@@ -106,15 +106,12 @@ test('CI runs the gate on every push and pull request, Node 22.13 and 24, on Lin
assert.doesNotMatch(job('linux'), /\n {4}if:/, 'Linux runs every time');
});
-test("the README shows the workflow's status badge [C7]", () => {
+test("the README header shows the board's own badges, not the CI status badge [C7]", () => {
+ // The person, 22 Oct: drop the gate badge from the header; a live red CI badge says less than the board's verified record.
const readme = readFileSync(resolve(ROOT, 'README.md'), 'utf8');
const section = readme.search(/^## /mu);
assert.notEqual(section, -1, 'the README has a section after its header');
const header = readme.slice(0, section);
- const workflow = 'https://github\\.com/pullboard-dev/pullboard/actions/workflows/gate\\.yml';
- const githubBadge = 'https://github\\.com/pullboard-dev/pullboard/actions/workflows/gate\\.yml/badge\\.svg';
- const shieldsBadge = 'https://img\\.shields\\.io/github/actions/workflow/status/pullboard-dev/pullboard/gate\\.yml';
- const markdown = new RegExp(`\\[!\\[[^\\]]*\\]\\((?:${githubBadge}|${shieldsBadge}(?:\\?[^)]*)?)\\)\\]\\(${workflow}\\)`, 'u');
- const html = new RegExp(`<a\\b(?=[^>]*\\bhref="${workflow}")[^>]*>\\s*<img\\b(?=[^>]*\\balt="gate")(?=[^>]*\\bsrc="(?:${githubBadge}|${shieldsBadge}(?:\\?[^\"]*)?)")[^>]*>\\s*<\\/a>`, 'u');
- assert.ok(markdown.test(header) || html.test(header), 'the header links a GitHub or shields status badge to the gate workflow');
+ assert.doesNotMatch(header, /actions\/workflows\/gate\.yml/u, 'the header carries no CI status badge');
+ for (const badge of ['pullboard', 'verified', 'npm']) assert.match(header, new RegExp(`alt="${badge}(?:: [^"]+)?"`, 'u'), `the header shows the ${badge} badge`);
});
diff --git a/test/readme.test.js b/test/readme.test.js
index a32dba24..5460ebe4 100644
--- a/test/readme.test.js
+++ b/test/readme.test.js
@@ -100,8 +100,8 @@ function wordCount(text) {
test('README has the approved centered header, logo and one-line facts [I10,I12]', () => {
assert.match(README, /^<p align="center">\n <picture>\n <source media="\(prefers-color-scheme: dark\)" srcset="docs\/img\/logo-dark\.svg">\n <img src="docs\/img\/logo\.svg" alt="" width="72">\n <\/picture>\n<\/p>\n\n<h1 align="center">Pullboard<\/h1>/u);
- assert.match(README, /<p align="center"><a href="https:\/\/pullboard\.dev"><b>pullboard\.dev<\/b><\/a> · lives in your git repo · no account · no dependencies · never calls a model<\/p>/u, 'the centered header pins the product facts in one line');
- for (const badge of ['gate', 'npm', 'node', 'MIT license']) assert.match(README, new RegExp(`alt="${badge}"`), `${badge} badge remains in the header`);
+ assert.match(README, /<p align="center"><a href="https:\/\/pullboard\.dev"><b>pullboard\.dev<\/b><\/a> · lives in your git repo · no dependencies · <b>Walk away, and it builds\.<\/b><\/p>/u, 'the centered header pins the product facts and the tagline in one line');
+ for (const badge of ['pullboard', 'verified', 'npm', 'MIT license']) assert.match(README, new RegExp(`alt="${badge}(?:: [^"]+)?"`), `${badge} badge remains in the header`);
assert.ok(readFileSync(join(ROOT, 'docs/img/logo.svg'), 'utf8').startsWith('<svg '), 'the approved light logo is present');
assert.ok(readFileSync(join(ROOT, 'docs/img/logo-dark.svg'), 'utf8').startsWith('<svg '), 'the approved dark logo is present');
});
@@ -307,9 +307,9 @@ test('README pins its approved Philosophy, Key concepts and Why structure [I10,I
const whyStart = README.indexOf('### Why', concepts);
const nextTopLevel = README.indexOf('\n## ', philosophy + 1);
assert.ok(philosophy >= 0 && philosophy < concepts && concepts < whyStart && whyStart < nextTopLevel, 'Philosophy contains Key concepts and Why before the next top-level section');
- const approvedOpening = "**Accumulated judgement for agents.** Agents are good at tasks and bad at projects: every new session starts without what you decided, so you become the glue. The agents scale; your judgement doesn't. Pullboard is a structural approach to orchestrating agents coherently over long-term builds, so you can trust your agentic workforce to fill in the blanks the way you would, without re-prompting them.";
+ const approvedOpening = 'Pullboard keeps agents building **true to your intent**, with a second agent checking every change. Your judgement carries across every agent and every session, so a complex project stays coherent for months. It\'s a coordination queue for serious software: you decide what matters once, and many agents build, check and land the work.';
const openingWithFigure = `${approvedOpening}\n\n${figureMarkup('docs/img/holds.svg', 'How it holds. Your decision becomes a bar each piece of work is frozen against; a second agent checks the exact commit; the board keeps what passed, what failed and why; the next agent starts from there.')}\n\n**Vibe code a real product.**`;
- assert.ok(README.includes(openingWithFigure), 'the approved opening and holds figure sit directly before the unchanged lead');
+ assert.ok(README.includes(openingWithFigure), 'the approved opening, then the holds figure, then the lead (the person, 22 Oct)');
assert.equal(README.split(approvedOpening).length - 1, 1, 'the approved opening appears exactly once');
assert.doesNotMatch(`${approvedOpening} ${FIGURES.holds()}`, /\b(?:memory|remember)\b/iu, 'the opening and its figure avoid those terms');
const section = README.slice(philosophy, nextTopLevel);