Install Options

Every install path, for every tool. The shortest one is the Quick Start; this page is the rest: per-adapter flags, what each writes and where, the plugin and LSP paths, and how to install more than one adapter.

# 1. Clone the latest stable release (drop --branch for unreleased master)
git clone --branch v1.1.0 https://github.com/gosha70/code-copilot-team.git
cd code-copilot-team

# 2. Install for your tool(s)
./scripts/setup.sh --claude-code                    # Claude Code → ~/.claude/
./scripts/setup.sh --pi                             # Pi (enforced) → ~/.code-copilot-team/pi/
./scripts/setup.sh --codex                          # OpenAI Codex → ~/.codex/
./scripts/setup.sh --cursor ~/my-project            # Cursor → project/.cursor/
./scripts/setup.sh --github-copilot ~/my-project    # GH Copilot → project/.github/
./scripts/setup.sh --windsurf ~/my-project          # Windsurf → project/.windsurf/
./scripts/setup.sh --aider ~/my-project             # Aider → project/CONVENTIONS.md

# Or install everything at once
./scripts/setup.sh --all ~/my-project

# Re-sync after pulling repo updates
git pull && ./scripts/setup.sh --sync --claude-code

Reading the docs on the web

The same documentation is published as a searchable site at https://gosha70.github.io/code-copilot-team/, built from master on every change to the docs. The guides are equally readable here in the repository and in the Studio’s Learn tab — all three serve the same files.

What each adapter writes

Tool	Adapter output	Install location
Claude Code	agents, hooks, commands, settings	~/.claude/ (global)
Pi	enforcement runtime extension + skills/prompts	pi install (advisory) / pi-code (enforced)
OpenAI Codex	AGENTS.md + 5 skills	~/.codex/ (global)
Cursor	.mdc files with frontmatter	project/.cursor/rules/
GitHub Copilot	copilot-instructions.md + per-rule instructions	project/.github/
Windsurf	rules.md	project/.windsurf/rules/
Aider	CONVENTIONS.md	project/

Which of them can enforce the contract rather than merely read it is in the README’s Choose your tool table, generated from the feature catalog.

The legacy ./claude_code/claude-setup.sh path still works — it delegates to the adapter.

After git pull, run --sync to regenerate configs and re-install.

Alternative: Install as a Claude Code Plugin

For Claude Code users who prefer the plugin system over setup.sh:

# Add the CCT marketplace (one-time)
/plugin marketplace add gosha70/code-copilot-team

# Install the plugin
/plugin install code-copilot-team@code-copilot-team

The plugin is the quick path; setup.sh is the full harness. The plugin’s contents are generated from the same sources setup.sh installs (scripts/generate.sh), so the two never drift apart. Update installed plugins with /plugin marketplace update.

What the plugin installs

• All 24 skills, the 14 agents and the 14 commands, plus the SDD templates and the review-decide helper the commands and agents read.
• The same seven hooks as setup.sh: file protection, git safety, auto-format, type verification, test-on-stop, context re-injection, notifications.
• Plugin components are namespaced: commands are /code-copilot-team:shape, agents are code-copilot-team:build. Asking for “the build agent” by its plain name still reaches it.
• Commands that read a template read it from the plugin’s own directory, which is outside your project, so Claude Code’s usual permission check for reads outside the project applies.

What only setup.sh gives you

• Always-loaded rules. A plugin cannot ship instructions that load into every session. The four skills setup.sh installs as always-on rules (coding-standards, copilot-conventions, copyright-headers, safety) ship in the plugin as ordinary skills: they load when relevant, not always. origin-confirmation and wiki-first-query are not always-on under either install — they are on-demand skills, reached through the pointer in the global CLAUDE.md. The two protect hooks enforce file and git safety either way.
• The global CLAUDE.md manifest, the claude-code launcher, the status line, and the templates for new projects.
• The peer-review and memkernel hooks — those are CCT-pipeline-specific.

Using both

Both install paths coexist. Claude Code does not merge same-named components: you will see /shape and /code-copilot-team:shape, build and code-copilot-team:build, side by side, and they behave the same. The cost is that every skill and agent is listed twice in context (about 1,500 extra tokens). The plugin’s copies of the five non-safety hooks step aside when setup.sh’s copies are installed, so formatting, verification and notifications run once. The two protect hooks never step aside: with both installs they run twice, which costs a duplicate block message and nothing else.

Recommended: Install LSP Plugins (Claude Code)

For continuous type-error feedback during edits, install the appropriate code-intelligence plugin. Each requires its language-server binary on $PATH:

# Install the language server first, then the plugin:
pip install pyright && /plugin install pyright-lsp@claude-plugins-official           # Python
npm i -g typescript-language-server typescript && /plugin install typescript-lsp@claude-plugins-official  # TypeScript
go install golang.org/x/tools/gopls@latest && /plugin install gopls-lsp@claude-plugins-official  # Go

These provide native LSP diagnostics and are preferred over the bundled verify-after-edit.sh hook. The hook remains as a fallback for languages without an LSP plugin. See the official plugin catalog for all available languages.