Shared team knowledge base template. Designed to be forked and customized for your team's domain. Consumed by team members via git submodule inside their personal ContextOS Personal.
Last Updated: 2026-04-14
Every team has hard-won knowledge scattered across wikis, Slack threads, shared drives, and people's heads. When someone needs to do a common task, they either reinvent the wheel or hunt down the person who's done it before.
This repo is your team's single source of truth for AI-assisted work. It gives every team member access to the same reference docs, tool documentation, vetted prompt templates, guardrails, and step-by-step workflows — loaded directly into Claude Code as context.
When a senior team member builds a great prompt, it goes here. When processes change, it gets updated once — and the whole team has it. New hires get the team's collective expertise on day one.
- Core files at the root (
team-overview.md,tools-and-integrations.md,guardrails.md,CLAUDE.md,CONTRIBUTING.md) are domain-neutral template shells. They ship with{{PLACEHOLDER}}strings that the team lead fills during setup. - Full reference examples live in
examples/(currentlyexamples/infosec/— a complete InfoSec team context: prompts, workflows, integrations, compliance frameworks, security guardrails). - Your team's customized content replaces placeholders in the core files and can expand into
workflows/,prompts/,integrations/as needed.
This way you get a blank slate that respects the pattern, with a fully-filled example to learn from.
Team members don't clone this repo directly. It's included as a git submodule inside each person's ContextOS Personal:
~/context-os/
├── CLAUDE.md ← Personal router (auto-loaded by Claude Code)
├── my-context.md ← Personal details
└── shared-context/ ← THIS REPO (git submodule)
├── team-overview.md
├── tools-and-integrations.md
├── guardrails.md
└── workflows/...
Claude Code reads shared-context/ like any local file. No special config needed.
- Fork this repo to your organization.
- Clone your fork:
git clone https://github.com/YOUR-ORG/contextOS-team.git cd contextOS-team - Run Claude Code:
A SessionStart hook detects the uncustomized state (placeholders in core files) and triggers skills/team-setup.md. Claude walks you through an 8-step flow covering team structure, tools, standards, SLAs, guardrails, and prompt priorities (~15 min).
claude
- Review, commit, push:
git diff git add -A && git commit -m "Customize team context for [Your Team]" git push origin main
- Enable branch protection — see CONTRIBUTING.md.
- Point
contextOS-personalat your fork. In your team's version ofcontextOS-personal, update.gitmodules:[submodule "shared-context"] path = shared-context url = https://github.com/YOUR-ORG/contextOS-team.git
Your team members get your customized content automatically.
| File | Purpose |
|---|---|
CLAUDE.md |
Auto-loaded by Claude when working in this repo; routes maintenance tasks |
team-overview.md |
Team structure, roster, escalation, SLAs, cadence |
tools-and-integrations.md |
Tool inventory and integration docs |
guardrails.md |
AI usage rules (data classification, restricted actions, approvals) |
CONTRIBUTING.md |
How team members propose changes (PR process, review matrix) |
skills/team-setup.md |
Onboarding flow for team leads |
As the team grows the knowledge base, expect to create:
workflows/— step-by-step procedures for recurring tasksprompts/— vetted prompt templates, typically role-specificintegrations/— how Claude connects to specific tools your team usesstandards.md— industry frameworks, coding standards, brand guidelines, etc.
A complete InfoSec team context: 20+ approved prompts, 5 workflows, 7 integration guides, compliance frameworks (FedRAMP, CMMC, NIST, SOC 2), role-specific prompts for SOC analysts, cloud engineers, compliance managers, incident responders, risk/GRC analysts.
Use it as a reference when building your team's own. Copy files directly if they fit your domain.
# Pull latest team knowledge (one command)
cd ~/context-os
git submodule update --remoteWhen someone PRs an update, every team member gets it with that one command.
This structure fits any team. Possible forks:
- Engineering — architecture docs, coding standards, deployment workflows, on-call runbooks
- InfoSec — security frameworks, SIEM/EDR tooling, compliance, SOC runbooks (fully built in
examples/infosec/) - Marketing — brand guidelines, campaign workflows, content review prompts
- Legal — contract review workflows, compliance checklists, clause libraries
- Finance — reporting workflows, audit procedures, forecasting templates
- Operations — process documentation, vendor management, SOP library
- Data science — model review checklists, evaluation patterns, reproducibility guides
- Healthcare — clinical workflow references, regulatory constraints, HIPAA-safe AI usage
- Team-wide knowledge (architecture, tools, procedures)
- Industry framework references relevant to your domain
- Vetted prompt templates that produce consistent results
- Workflows anyone on the team would use
- AI usage guardrails and constraints
- Personal context or preferences
- Credentials, API keys, or sensitive data
- Company-specific details that shouldn't be public
- Unverified or speculative content
See CONTRIBUTING.md for the full PR process.
MIT