agentleFS

All solutions / Shared knowledge

Stop rebuilding context in every tool.

Keep project knowledge and playbooks in one place. Teammates can change tools while their agents keep working from the same sources.

Your second brain, online for every AI you use.

Bring your notes, linked documents and project knowledge into one shared workspace. Teammates and their agents can work from the same context across different platforms, within the access you give them. Add a GitHub repo or a Drive folder to follow its source.

Your second brain, reachable from many clients and a collaborator A graph of linked notes sits at the centre inside a thin blue circle, your vault online. Grey spokes reach teammates’ AI clients, coding and review agents, different harnesses, and any MCP client. A few notes at the lower left are outlined in purple; a collaborator, in purple, reaches only those: what you shared. [[Ideas]] Your AI Teammate’s AI Review agent Another harness Background agent Coding agent any MCP client Uploaded once. backlinks keep working your vault, online over MCP a collaborator only what you share

Swipe to see the whole diagram →

Shared skills and playbooks for every agent.

Keep your team’s skills, instructions and ways of working together. Teammates can use different models and harnesses while drawing on the same playbooks, with each agent told what applies to its work.

Told what applies where it works A folder tree. company/ holds AGENTS.md; research/ holds a public skill, citations; site/ holds AGENTS.md and your skill design-review; site/launch/ holds a public skill, brand-voice, copied from acme/brand-kit; site/blog/ holds its own AGENTS.md. Your agent works in site/launch/. An orange path runs from it up to the root. Its scope is site/, drawn as a dashed box, so company/AGENTS.md lies outside it and is not named. It is told site/AGENTS.md, design-review and brand-voice, and not citations or blog's AGENTS.md. Blue marks your files, grey marks public ones. your agent’s scope: site/ outside its scope: not named company/ AGENTS.md research/ .claude/skills/citationspublic site/ AGENTS.md .claude/skills/design-review launch/ .claude/skills/brand-voicepublicfromacme/brand-kit blog/ AGENTS.md your agent Your agent in launch/ is told 3 of the 6, nothing else site/AGENTS.md design-review brand-voice It reads each one when it needs it. yourspublic

Swipe to see the whole diagram →

A public library of skills and agent instructions.

Skills and agent instruction files from public GitHub repositories, kept current and searchable. Under each one, people and their agents say what worked and what didn't, so the next one starts from that.

Browse them all, no account needed

Read the document and folder guide

Connect your harness