A shared team knowledge mounting system for AI coding agents. Fork this template, fill it with your team's knowledge, and let your AI agent reference it automatically.
AI coding agents start every conversation from scratch. They don't know your team's conventions, glossary, project structure, or who owns what. Team Context fixes this — mount a shared knowledge repo and your agent instantly understands the team.
┌─────────────────────────────────┐
│ Team repo (GitHub/GHE) │
│ conventions/ glossary/ ... │
└──────────────┬──────────────────┘
│ mount
▼
┌─────────────────────────────────┐
│ ~/.claude/team-context/{id}/ │
│ Local cache + CLAUDE.md inject │
└──────────────┬──────────────────┘
│ auto-reference
▼
┌─────────────────────────────────┐
│ AI agent reads team knowledge │
│ during any conversation │
└─────────────────────────────────┘
- Fork this template repo
- Fill it with your team's docs (conventions, glossary, projects, services)
- Mount it in Claude Code:
Install team context from: https://[your-pages-url]/SETUP.md - Your agent now knows your team's knowledge
| Directory | Purpose |
|---|---|
glossary/ |
Team-specific terms, acronyms, internal system names |
conventions/ |
Coding conventions, naming rules, branch strategy |
projects/ |
Current projects, epics, sprint info |
services/ |
Service architecture, dependencies, ownership |
docs/ |
Onboarding guides, process docs |
| Command | What It Does |
|---|---|
| mount | Install a team context from a SETUP.md URL |
| unmount | Remove a mounted team context |
| update | Pull latest from remote repo |
| push | Add/edit content and push back to the shared repo |
Add .context/sources.yml to auto-sync from Confluence, Jira, or GitHub:
sources:
- id: team-wiki
type: confluence
space: YOUR_SPACE_KEY
sync_to: docs/See sources.example.yml for the full template.
Built-in safety rules prevent collecting secrets, PII, personal opinions, or org-sensitive information. The guardrails enforce: improve systems and processes, never judge people.
MIT