yc
Skill4 starsChanged 2 days agoMITView on GitHub
Draft or refresh a Y Combinator application from a founder hub's real evidence. Confirms the current batch, deadline and application questions on the web, answers only from logged evidence, marks unsupported claims, and lists the evidence to get before submitting. Use when preparing a YC application, a founder video script, or any accelerator or investor form.
In this repository50 other files it gives its agents.
- agent-writingSkill
- arenaSkill
- blast-radiusSkill
- bun-playwright-runner-selectionSkill
- cinematic-shotSkill
- clean-ui-shotSkill
- codebase-designSkill
- debug-loopSkill
- dep-upgradeSkill
- diagnosing-bugsSkill
- discoverSkill
- domain-modelingSkill
- earshotSkill
- eval-judgeSkill
- flaky-testSkill
- grillingSkill
- grillSkill
- idea-scoutSkill
- improve-codebase-architectureSkill
- interrogateSkill
- make-mediaSkill
- migrate-to-rustSkill
- migration-planSkill
- multi-agent-windows-hazardsSkill
- perf-loopSkill
- postmortemSkill
- prototypeSkill
- prSkill
- real-soundSkill
- recallSkill
- release-notesSkill
- repo-mapSkill
- researchSkill
- retroSkill
- secure-diffSkill
- sellSkill
- subagent-report-recoverySkill
- tasteSkill
- taste-skillSkill
- tddSkill
- teachSkill
- to-questionnaireSkill
- to-specSkill
- to-ticketsSkill
- wayfinderSkill
- weeklySkill
- windows-bash-backslash-halvingSkill
- wizardSkill
- writing-for-agentsSkill
- ystack-upgradeSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
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
- Very specific: line 35
Read bottleneck.md, ideas.md, the committed idea's evidence files and founder.md if present.
- Very specific: line 66
Every claim traces to a file in the hub.
≈ “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.
Files
/yc: an application built on evidence
The founder hub
This skill reads and writes a founder hub: one folder holding the startup's truth.
- Location: the
FOUNDER_HUBenvironment variable, default~/founder. Below,<hub>means that folder. - Files this skill uses:
| File | What it holds |
<hub>/bottleneck.md | The current bet, this week's single bottleneck, the goal and its deadline. This skill writes the application deadline here |
<hub>/ideas.md | Candidate ideas and their status; the committed one is what the application is about |
<hub>/evidence/<idea>.md (+ -prospects.md, -pipeline.md) | Conversations, pilots, users, payments, metrics, with dates |
<hub>/founder.md (optional) | The founders' background and proof of shipping and selling, in their own words |
<hub>/yc/<batch>-draft.md | The application drafts this skill writes |
If the hub or a file is missing, say so and offer to create the minimal structure: the folder, an
empty bottleneck.md (current bet, bottleneck, goal, deadline), ideas.md with a table header (idea,
one line, who pays, evidence so far, status), an empty evidence/ and yc/. With no evidence yet,
still draft, but most answers will be [NEEDS EVIDENCE], and say plainly that the application isn't
ready. Never fail silently.
Read bottleneck.md, ideas.md, the committed idea's evidence files and founder.md if present. If no
idea is committed, say so and draft for the strongest candidate, clearly labelled as such.
Steps
- Confirm the facts on the web (web search and fetch, starting at ycombinator.com/apply): the
current or next open batch, its application deadline, and the current application questions. Never
assume a batch, a date or a question list from memory; they change every cycle. If the user named a
batch, check it is still open. Record the source URL and the date you checked, and write the
deadline into
bottleneck.md. - Draft
yc/<batch>-draft.md, answering every current question in plain, specific, short sentences. YC readers value clarity and concrete numbers.- The one-liner: what it is, for whom, the outcome.
- Progress and traction: only what
evidence/supports (conversations held, pilots, users, revenue, growth), with dates. Never invent or round up a number. If there's nothing yet, say what's being done this week to get some. - Why now, the insight, competitors: be specific. Name real competitors and say what you know that they don't.
- The founders: the real story, including concrete proof of shipping and selling (from
founder.mdor what the user tells you). Ask rather than embellish.
gaps: the 3 to 5 pieces of evidence that would most strengthen the application before the deadline, each with the action that gets it (usually through thediscoverorsellskills, if installed).video: a 60-second founder video script: who you are, what you're building, the evidence, why you.refresh: re-check step 1, then update the existing draft from evidence logged since, and list what changed.
Rules
- Evidence over adjectives; no hype words.
- Every claim traces to a file in the hub. Mark anything unsupported as
[NEEDS EVIDENCE]. - Draft only. The founders review and submit the application themselves; never submit a form or send anything on their behalf.
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
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.

