create-vivary 0.2.7

Scaffold a complete Vivary agent workspace — the create-t3-app for agent-native workspaces.

create-vivary

Scaffold a complete Vivary agent workspace: tropo config, strato workspace files, runtime skills, private-memory boundaries, progressive module indexes, and a starter typed graph.

Current release: 0.2.7. Use 0.2.7 for new installs; no migration is expected from 0.2.1, 0.2.2, 0.2.3, 0.2.5, or 0.2.6.

Release focus: 0.2.7 ships the context-compression release surface: generated active-context guidance now leads with tropo find, points agents to the canonical LLM active-context guide, and names LanceDB as embedded storage rather than hidden search behavior.

Security hardening: The 0.2.5 line validates active .gitignore rules for USER.md, MEMORY.md, memory/*, and heartbeat-reports/*; scaffolds private heartbeat report storage; and refuses symlinked or out-of-workspace scaffold, storage, and cleanup paths.

Install & scaffold

pip install create-vivary==0.2.7              # or: uvx create-vivary@0.2.7 ...
create-vivary my-workspace --preset coding    # interactive wizard on a TTY
create-vivary my-workbench --preset knowledge-work --memory local
create-vivary my-codebase --preset coding --active-context cocoindex-code
create-vivary capabilities --preset second-brain --json
create-vivary doctor my-workspace

# Agent-mode (no prompts, machine-readable output):
create-vivary init . --preset coding --auto --size large --yes --json

# Reconfigure storage on an existing workspace:
create-vivary wizard my-workspace --auto --storage embedded --yes --json

create-vivary <name> is shorthand for create-vivary init <name>; pass init / doctor / wizard / capabilities explicitly whenever you prefer. The same UX is available on npm via the @vivary/create launcher (npm create @vivary@latest my-workspace), versioned in lockstep.

Presets share the same agent OS shell, then seed a different starter graph. Each starter module is generated as modules/<id>/index.md so agents route through a small module index before opening deeper context:

Preset	Module	First slice	Verification
coding	codebase	local-ci-baseline	local-checks
second-brain	knowledge-base	capture-routine	retrieval-smoke
knowledge-work	workbench + sources	workbench-first-artifact	workbench-proof
writing	manuscript-system	draft-review-loop	editorial-review

The command is local-only. With --storage embedded or --auto on a large workspace, it self-installs vivary-tropo[embedded] (LanceDB) with a confirmation prompt, or silently with --yes. Scaffold writes, storage config writes, and stale generated cleanup refuse symlinked destination parents, including when --force is used, so output stays inside the selected target. Use --dry-run to simulate without writing, installing, or cleaning stale files. For scripted storage selection, pass --no-wizard --storage embedded --yes or use --auto; in human mode, the wizard asks and its answers drive storage.

Semantic memory is a separate optional capability. --memory local writes local-only policy and graph nodes; --memory cognee writes Cognee policy and verification docs. Neither option indexes content or sends data anywhere during scaffold, and Cognee is not installed by default. Use create-vivary capabilities --preset <name> --json to show agents the available optional pieces before setup.

For coding workspaces, --active-context cocoindex-code adds an optional CocoIndex-code sidecar profile: active-context skills for Claude/Codex-style agents, local policy docs, graph nodes, and .cocoindex_code/ in .gitignore. It does not auto-install CocoIndex-code, create an index, or enable MCP; the generated docs give the approved ccc init / ccc index path, and the skill points agents to the canonical copyable LLM guide.

doctor validates the generated shell, active privacy ignore rules, module directory indexes, semantic-memory status, and typed graph:

python packages/create-vivary/create_vivary.py doctor sandboxes/coding-demo --json

Developing from source

python packages/create-vivary/create_vivary.py init sandboxes/coding-demo --preset coding
python packages/create-vivary/create_vivary.py doctor sandboxes/coding-demo
python packages/tropo/tropo.py check --root sandboxes/coding-demo

---

Website & docs: https://vivary.vercel.app/