LinkedIn .md context files
Attach these once to a Claude Project, a ChatGPT Project or an agent folder, and stop re-explaining your business on every request. A prompt is consumed by one message. A context file is read on every turn, for months — which is why these are versioned, structured and filled with worked example content rather than shipped empty.
Fifteen files, and you probably need four
This is the full set. In practice a working folder has ICP.md, PRODUCT.md, OUTREACH.md and LINKEDIN-LIMITS.md filled in, and the rest added as the work needs them.
linkedin-ai/
├── CLAUDE.md # the root instruction — read first
├── AGENTS.md # roles when you run more than one
├── LINKEDIN-LIMITS.md # the real ceilings. maintained by us
├── ICP.md # start here
├── PERSONAS.md
├── COMPANY.md
├── PRODUCT.md
├── BRAND-VOICE.md # five real sentences
├── OUTREACH.md
├── OBJECTIONS.md
├── RESEARCH.md
├── SALES-NAVIGATOR.md
├── CONTENT-ENGINE.md
├── RECRUITER.md
└── JOB-SEARCH.mdEvery answer starts already knowing your business
This is the mechanism, not a description of it: the same question asked with no files attached gets a template; asked inside a Project holding these four, it gets an answer checked against your actual ICP and your actual send limit.
- ICP.md
- PRODUCT.md
- OUTREACH.md
- LINKEDIN-LIMITS.md
Write a first message to a VP Finance I found through Sales Navigator.
Checked against ICP.md — 400-person logistics company, multi-entity finance. In scope.
Checked against LINKEDIN-LIMITS.md — this is invitation #34 this week for this seat, inside the ~100 ceiling. Sending it does not push you over.
Draft, 214 characters — a message built from OUTREACH.md's field limits and banned openers, not a generic template.
Every file, free to read and download
- Context file
ICP.md
Every prompt you run afterwards knows who you sell to, in terms LinkedIn can filter on.
- Context file
LINKEDIN-LIMITS.md
Your assistant stops proposing 300 connection requests a week, because it now knows the ceiling is about 100.
- Context file
BRAND-VOICE.md
Drafts arrive in your voice on the first attempt instead of after four rounds of "less corporate".
- Context file
COMPANY.md
The assistant stops describing your company the way its training data describes companies in your category.
- Context file
PRODUCT.md
Messages stop describing features and start describing what changes for the person receiving them.
- Context file
OUTREACH.md
Every message an assistant writes comes out at the right length, in the right shape, with the openers you hate already banned.
- Context file
OBJECTIONS.md
The assistant answers the objection you actually get, in the way that has worked, instead of improvising.
- Context file
CLAUDE.md
An agent that reads the right files in the right order and knows what it is not allowed to do.
- Context file
AGENTS.md
Three narrow agents with defined handoffs instead of one conversation that drifts.
- Context file
RESEARCH.md
Research that is consistent between people and between runs, with the evidence line drawn in the same place every time.
- Context file
SALES-NAVIGATOR.md
Search instructions that reference filters your tier actually has.
- Context file
PERSONAS.md
Messages that speak to what the reader is measured on rather than to a job title.
- Context file
CONTENT-ENGINE.md
A posting system with a material pipeline, rather than a calendar that dies in week three.
- Context file
RECRUITER.md
Sourcing and outreach that stays inside your InMail budget and does not infer things it should not.
- Context file
JOB-SEARCH.md
Every draft — message, application, profile rewrite — argues the same case about you.
Questions about context files
What is a context file?
A markdown file you attach once to a Claude Project, a ChatGPT Project, a Gemini Gem or an agent folder, and which the assistant then reads on every turn. The practical difference from a prompt: a prompt is consumed by one message and dies with the conversation, while a context file persists for months. It is why your tenth request in a Project is better than your first, and why a good file set removes most of the repetition from working with an assistant.
Why not just paste this into the chat each time?
Because you will paste it slightly differently every time, and no two outputs will be built on the same definition of your ICP. The value is not that the text is in context — it is that the same text is in context, unchanged, across every conversation for a year.
Do these work outside Claude?
Yes. ChatGPT Projects take uploaded files, Gemini Gems take an instruction plus attachments, and Cursor and Claude Code read a CLAUDE.md at the root of a workspace automatically. The files are plain markdown with no tool-specific syntax, which is deliberate.
Which file should I fill in first?
ICP.md, unless you are here for content, in which case BRAND-VOICE.md. Both are the input to everything downstream, and a folder where those two are done properly beats one where all fifteen are half done.
Take all fifteen as one zip
The Starter Pack is every context file plus the twelve most-used prompts, assembled into a folder. It costs an email, and the ungated link is under the form.