Set up team-ops for this machine/project - create .claude/team-context.json, scaffold the team-shared ai/ folder skeleton, and optionally register a project. Use when onboarding a machine to a team shared folder, when the team-context hook reports a config error, or when adding a new team project.
Installs into .claude/skills of the current project.
Are you the author of Setup Team Context?
Add the live security badge to your README. It updates with every re-scan.
[](https://www.skillsdirectory.com/skills/atman-33-setup-team-context)
---
name: setup-team-context
description: Set up team-ops for this machine/project - create .claude/team-context.json, scaffold the team-shared ai/ folder skeleton, and optionally register a project. Use when onboarding a machine to a team shared folder, when the team-context hook reports a config error, or when adding a new team project.
---
# Set up the team-ops context
Configure this working directory to use a team-shared folder as the team's
single source of truth (see `${CLAUDE_PLUGIN_ROOT}/docs/design.html`).
## Steps
1. **Collect inputs** (ask only for what's missing):
- `teamRootPath` — the locally synced shared folder (Google Drive /
OneDrive / file server). The `ai/` zone is created inside it.
- `me` — the user's short name (used in activity-log entries).
- Optionally a project name to scaffold, and the team content language
(BCP 47 tag like `ja` / `en`; only applied when `_meta/team.json` does
not exist yet).
2. **Run the setup script**:
```sh
node "${CLAUDE_PLUGIN_ROOT}/scripts/setup/init-team-context.mjs" \
--team-root "<path>" --me "<name>" [--project "<name>"] [--language ja]
```
It writes/merges `.claude/team-context.json` and keeps it out of git
through `.git/info/exclude` — never `.gitignore`, which in a workhub vault
is rewritten by every template update. If the output carries a `warning`
that the file is already tracked, run the `git rm --cached` it names. It
scaffolds `ai/` (`knowledge/`, `_meta/conventions.md`, `_meta/team.json`,
`_meta/activity-log.md`), and — with `--project` — the project skeleton
(`config/project.json`, `backlog/`, `sprints/`, `docs/spec/spec.md`,
`reports/daily/`). It only creates missing files, never overwrites.
3. **Configure the project's repos** (when a project was scaffolded): open
`projects/<name>/config/project.json` and replace the example entry with
the real repositories — `name`, `url`, and each repo's `devMainBranch`
(the project development main branch PBI work merges into).
4. **Report** the JSON summary and remind the user that the new context is
injected on the next session start.
## Notes
- The local config holds machine-local paths only; everything shared lives
in the shared folder itself, so teammates just point at the same
`teamRootPath`.
- Changing the team language later = editing `_meta/team.json`.
