Skip to content
Sign up

The skill sharing platform for agentic engineering teams

One home for your team’s agent skills. Synced to every developer, scoped to each session, traced end to end.

Get started in 5 minBook a demo
$ brew install to11ai/tap/to11
claude-code
$ to11 skill sync
home   ~/.claude/skills
  + incident-triage    v4   sre/runbooks
  ~ house-style        v8 → v9
  + release-checklist  v3   web/checklists
2 installed, 1 updated, 6 in sync

01The problem

The AI labs neglected how teams work when creating skills.

Teams accumulate procedures worth standardizing with skills. But there’s no simple path from their laptop to others.

  • 001

    Staleness

    A wrong procedure is worse than no procedure. Remove old skill versions.

  • 002

    Noise

    Essential in one repo, clutter in the next. Focus the agent’s attention.

  • 003

    Divergence

    No canonical home means no “latest”. Five people, five variants, none of them authoritative.

  • 004

    No link to the work

    The agent runs on a laptop the platform never sees. Connect a skill’s use back to the work it produced.

02How it works

Trace, store, sync, improve.

  1. 01

    Trace

    See a full trace of your coding work and the skills each session used.

    to11 code claude
  2. 02

    Store

    Keep a new or improved skill private while you evaluate it.

    to11 skill store
  3. 03

    Sync

    Release it. Every developer gets it on their next sync.

    to11 skill release
  4. 04

    Improve

    See which skills help and which need refinement, then store the next version.

03Solution

Effortlessly apply the right skill at the right time.

Scope agent skills per terminal session. Load, drop, or swap them any time the task changes, mid-session, no restart required.

  • 03.1One repo, one procedure

    Scoped to the work in front of you

    • Get the skills published by your team.
    • Scope to what you’re working on, and turn off what you don’t want.
    • Keep one set of skills across everything you work on.
    your repo · terminal
    # .to11/skills.yaml, committed to the repo
    projects:
      - project: web/checklists
        skills:
          release-checklist: { label: live }
          legacy-migration: { disabled: true }
    $ to11 skill sync
    repo   ./.claude/skills
      + release-checklist  v3   web/checklists
    1 installed, 4 in sync
    Example: the repo's own skills, one turned off

  • 03.2Hold a version through a release

    Current by default, pinned on purpose

    • Run the current skill version by default, without doing anything to get it.
    • Stay on a specific version when the current one doesn’t suit you.
    • Set a version rule for a whole group of skills in one statement. Break it for a single skill.
    to11 skill list
    projects:
      - project: sre/runbooks
        label: live
    $ to11 skill list
    SKILL           VERSION            STATE
    incident-triage live · 7           in-sync
    house-style     live · 9           in-sync
    postmortem      live · 2           in-sync
    Example: one skill held back

  • 03.3Fix a skill from a bad session

    Built around how skills actually get written

    • Update non-performing skills from any coding session.
    • Keep a new skill private to evaluate and publish it to your team when proven.
    • Analyse good and bad coding sessions and the skills used in each.
    terminal
    $ to11 skill store incident-triage --dir .
    stored    incident-triage v8  sre/runbooks
    installed ~/.claude/skills/incident-triage
    not published — `to11 skill release
      incident-triage` releases it on `live`
    $ to11 skill release incident-triage
    published incident-triage v8 → live
      sre/runbooks (home)
    everyone who follows `live` in that
    project receives it on their next sync
    Example: store privately, release when proven

  • 03.4Trace which skills helped

    Accountable, not invisible

    • Sync your coding sessions and see a full trace of your coding work.
    • See which skills help, and which need refinement.
    $ to11 code claude
Get started in 5 min
$ brew install to11ai/tap/to11

04Get started

New machine, day one.

One CLI. Every harness. Load exactly what your session needs, whatever agent’s got its hands on the terminal.

Anthropic’s CLI coding agent.

  1. 01

    Install the CLI

    $ brew install to11ai/tap/to11

    macOS and Linux, x86-64 and arm64.

    Other ways to install

    With asdf:

    $ asdf plugin add to11 https://github.com/to11ai/asdf-to11
    $ asdf install to11 latest
    $ asdf set --home to11 latest

    Or download a release archive and its checksums from GitHub releases.

  2. 02

    Set up this machine or this repository

    $ to11 ui

    Choose the project and label to follow and where skills land, for this machine or for a repository your whole team works in.

    Skills land in ~/.claude/skills

  3. 03

    Sync

    $ to11 skill sync

    Every skill your team has released arrives on the first sync. The same command keeps this machine current.

05Platform

Latest addition to the to11 platform.

The same platform that syncs your team’s skills also gateways your model calls, traces every run, and manages your prompts. One account, one CLI, one place your AI engineering actually lives.

06Impact

Skills to pay your bills.

Simplify your own life & help your colleagues along the way. Watch coding agent mastery improve across your org. Whatever you do load, crank that one to 11.

Pick your skillsBook a demo
$ brew install to11ai/tap/to11