brainbase

Pack, share, and install agent templates across harnesses (Claude Code, Codex, Kafka).

Install

npm install -g @brainbase-labs/cli

The package installs a brainbase command on your PATH. Requires Node.js 18+.

Quick start

brainbase login                 # connect this device to brainbase
brainbase template pack         # bundle the current agent into a template
brainbase template publish      # upload to the registry
brainbase template search       # find templates published by your team
brainbase template onboard <creator/slug>   # install or refresh a template

Run brainbase help to see every command.

Feature support by harness

Each component type is stored in the place that harness already reads from at runtime — the CLI doesn't impose its own layout. "Native" cells describe where the install lands; "—" means the harness does not support that component type.

| Component | Claude Code | Codex | Kafka | |---|---|---|---| | Skill | .claude/skills/<slug>/ (folder per skill, w/ provenance marker) | ~/.codex/skills/<slug>/ (global) or .agents/skills/<slug>/ (project) | .kafka/skills/<slug>/ | | MCP server | Project: .mcp.json. Global: ~/.claude.json → mcpServers. Never settings.json (Claude ignores it there). | Embedded in ~/.codex/config.toml as [mcp_servers.<slug>] blocks | Embedded in .kafka/kafka.json (no sidecar — the Kafka SDK reads MCPs straight from settings) | | Command (slash command / prompt) | .claude/commands/<slug>.md | ~/.codex/prompts/<slug>.md (Codex has no documented project-scope prompts dir) | .kafka/commands/<slug>.md (the Kafka runtime calls these "playbooks" internally, but the CLI keeps the command type for cross-harness symmetry) | | Sub-agent | .claude/agents/<slug>.md | — (Codex uses a different agent schema) | .kafka/agents/<slug>.md | | Playbook | .claude/playbooks/<slug>.md + auto-managed table at the bottom of CLAUDE.md | .codex/playbooks/<slug>.md + auto-managed table at the bottom of AGENTS.md | .kafka/playbooks/<slug>.md + auto-managed table at the bottom of KAFKA.md | | Instruction | Block in CLAUDE.md between <!-- brainbase:start name=<template> --> / :end markers | Same block convention in AGENTS.md | Same block convention in KAFKA.md | | Hook | — (todo) | — | — | | File | Arbitrary copy: single file or directory to a target path relative to cwd | Same | Same |

Playbook files carry a small YAML frontmatter (title, description) — the CLI parses it to build the table injected into each harness's instructions file. The table is bracketed by <!-- brainbase:playbooks:start --> / :end markers and regenerated on every pull, so removed playbooks drop out of the table automatically.

Project root (harness-agnostic)

• Auto-detected ONBOARDING.md and any extra files/folders you point --file <path> at — packed as file components and replayed at install time.

Configuration

| Variable | Purpose | |---|---| | BRAINBASE_REGISTRY_URL | Override the registry API host (default: https://kafka-llm-service.onrender.com) | | BRAINBASE_API_URL | Same as above; checked second | | BRAINBASE_TOKEN | Use a PAT instead of the stored login session | | BRAINBASE_HOME | Where local state lives (default: ~/.brainbase) |

Auth

Two ways to authenticate, in priority order:

1. brainbase login — interactive Supabase OAuth, stored in ~/.brainbase/auth.json. Carries full scopes.
2. brainbase token create --scope publish — long-lived PAT (bbpat_…) for CI / unattended use.

BRAINBASE_TOKEN env var overrides both.

License

MIT — see LICENSE.