Skills
Teach your AI once.
It remembers exactly when it matters.
Package any repeatable task as a Skill. Your AI team loads it automatically the moment a job calls for it. Build your own in plain language, or import skills from the open Agent Skills ecosystem.
Free to start. Scale as your library grows.
// what it is
A shared playbook binder for your AI.
A Skill is a reusable set of instructions, a "how we do X", that you create once and your AI agents load automatically when a matching task comes up. Assign it to a team, and every agent on that team can use it.
Think of it as a binder of saved how-tos. Each agent reaches for the right page exactly when a task needs it, then puts it back.
The problem it solves: people re-explain the same process to their AI over and over. They paste the same long prompt every time. Their AI team does not know the company's specific way of doing things. Skills fix that. Capture a repeatable task once, hand it to your team, and every agent runs it the same way without being re-taught.
Competitive teardown
WHEN · "TEAR DOWN" + COMPANY NAME
Pull the latest funding, leadership, top 3 products, our gaps, and the two angles we win on. Format: brief + threat-level + suggested next move.
Incident postmortem
WHEN · POSTMORTEM REQUESTED
Timeline → impact → root cause → contributing factors → action items with owners and due dates. Always blameless. Always pre-mortem next quarter's risks at the end.
Board update draft
WHEN · MONTHLY BOARD-UPDATE
Lead with the one number that matters this month. Highlights / lowlights / asks. Keep under 600 words. Match our previous-month tone exactly.
// how it works
Three steps, no code.
Create or import.
Write a skill in a few plain-language fields, a name, when to use it, the steps to follow. Or drop in a SKILL.md file in the open Agent Skills format.
Assign.
Turn it on for a specific AI team, or make it available across your whole workspace. The lead and every specialist on the team can use it.
Use it automatically.
Ask your AI something that matches and it loads the skill and follows it. Or just name the skill you want.
// example moments
What it looks like in real work.
SALES
"Run our competitive teardown on Acme."
The AI loads your Competitive Analysis skill and returns the structured output your team expects, every time.
ENGINEERING
"Draft the incident postmortem."
It follows your postmortem format every time, timeline, root cause, action items, the way your team writes them.
FOUNDERS
"Turn these notes into a board update."
Consistent and on-brand, every time. Pulls from your last board update's tone and structure.
SUPPORT
"Confirm the account before any change."
A support team carries a verification skill its agents pull up on the right conversations, automatically.
// why it stands out
Six things that matter.
01 · LOADED ON DEMAND
A library, not a wall of prompts.
Agents pull in a skill only when a task matches it. Build a full library, it never slows a conversation down or clutters it. Most "custom instructions" are always on; Skills arrive just in time.
02 · OPEN ECOSYSTEM
Import what you already have.
Paste or upload a SKILL.md file in the open Agent Skills format, the same files shared across the AI community. You are not starting from a blank page.
03 · TEAM-SHARED
The lead and every specialist.
Assign a skill to a zone (a team of AI agents grouped on your workspace map) and the whole team shares it. Or make it available everywhere in one click.
04 · NO CODE
Plain language. A few fields.
Author a skill the way you'd brief a teammate: a name, when to use it, and the steps to follow. That's it.
05 · BUILT FOR TRUST
Reviewed before they go live.
Imported skills are reviewed before they go live, and bundled scripts are not run. You stay in control of what your agents can do.
06 · FREE TO START
Build and import, always free.
Building and importing skills is free and unlimited. Paid tiers raise how many you keep active at once, with separate personal and team-workspace plans. Scale as your library grows.
// the open ecosystem
SKILL.md works here.
The Agent Skills format (SKILL.md) is an open convention shared across the AI community. If you have skills published in that format, from a teammate, an open-source repo, or a colleague's workspace, they import cleanly into In My Mind. Paste the text. Upload the file. Use it immediately. Skills you build here export the same way, so your library is portable.
// FAQ
Questions people ask first.
How is this different from a saved prompt or custom instructions?
A saved prompt is something you paste. Custom instructions are always on. A Skill is reusable across your whole team and loads only when it is relevant, so you can keep many of them without slowing anything down.
Will a big library slow my AI down?
No. Skills load only when a task matches, so the rest stay out of the way.
Can I use skills I already have?
Yes. Paste or upload a file in the open Agent Skills (SKILL.md) format.
Is there a cost?
Building and importing skills is always free, and you start with a set of active skills at no charge. Paid monthly tiers raise how many you keep active at once, with separate personal and team-workspace plans. Free to start, scale as you grow.
Who can create them?
You, for your personal workspace. A workspace owner or admin manages the shared library for a company.
Does it work with my whole AI team, not just one chat?
Yes. Assign a skill to a team and the lead and all of its specialists can use it.
// where Skills fits
Skills is one of four layers.
Repeatable work lives in pieces, a team's identity, its standing rules, its catalogue of how-tos, and the multi-step processes it runs on a clock. In My Mind has a layer for each. They are independent. They compose. Most real teams use all four together.
Who this team is and what it's for. A short charter that sits under everything else, the mission, the audience, the things this zone exists to do. Set once. Reads as the team's North Star. Every member sees it on every turn.
How this team always behaves. Numbered standing rules baked into every member's prompt. Tone of voice. Forbidden moves. The non-negotiables. If something must be true of every conversation in this zone, it goes here.
Specific ways to do specific tasks. A library of reusable how-tos your AI team pulls in only when relevant, competitive teardown, postmortem format, board-update draft. Build a hundred, none of them load until a task calls for one.
Whole multi-step processes you run. The thing that has steps, branches, approvals, and a schedule. Fires on cron, webhook, voice command, or by hand. Can hand work between zones, pause for human approval, and report when done.
All four, working together. Here is what a real sales-team zone looks like with every layer in play.
This team owns enterprise pipeline review and qualification. Always-on context for every member, every conversation.
Numbered rules: never approve a deal without the qualification checklist, always cite the CRM record, tone is direct, no hedging.
Twelve loaded in the library, competitive teardown, churn-risk triage, RFP first draft, deal-health summary. Each pulled in only when the task asks for it.
Three running on schedules, nightly deal-health digest (cron, sends to Slack), quarterly QBR pack (manual, fan-out to three sub-zones), incident postmortem (webhook trigger from PagerDuty).
Each layer can ship without the others. The four together are how a small team behaves like a large, well-trained one, every conversation, every time.
Give every agent your playbook.
Skills turn your best work into reusable expertise.
Free to start. Scale as your library grows.
Where Humans and AI collaborate.