NeuroCore Community is the public, contributor-friendly edition of NeuroCore. It ships the public package source, tests, docs, and contribution surfaces from the active development repo without the private operator state, local secrets, or internal-only working files.
The active architecture and scope contract lives in docs/ssd/.
Keep README-level guidance aligned with those documents when behavior changes.
Version: 0.1.0
Declared in pyproject.toml.
Main capabilities currently present in the public repository:
- capture notes, documents, and imported source material into durable memory
- query stored content with metadata filters and optional semantic ranking
- manage first-class brain manifests plus session and protocol workflows
- generate synthesized briefings and consensus-style report flows
- expose the same core behavior through library, CLI, HTTP, and MCP surfaces
- run SQLite, mirror, or Postgres-backed storage paths behind shared contracts
- validate repo metadata and scan for obvious secret-like values
- provide contribution surfaces for recipes, skills, integrations, dashboards, schemas, primitives, and curated extensions
The default bootstrap path is the mirror-first bootstrap script using the security operator profile:
python scripts/bootstrap.py
source .venv/bin/activateThat flow creates or reuses .venv, installs .[dev,semantic], writes an
operator-home env file under ~/.local/state/neurocore/, copies the local-only
config templates into that operator home, and runs pytest plus repo
validation unless you pass --skip-verify.
Local-only SQLite remains available as an explicit fallback when you do not want the default mirror path.
For a guided setup:
python scripts/bootstrap.py --wizardFor the checkout-safe CLI wrapper:
python scripts/neurocore_checkout.py diagnose
python scripts/neurocore_checkout.py capture --request-json '{"bucket":"research","content":"community repo note","content_format":"markdown","source_type":"note"}'
python scripts/neurocore_checkout.py query --request-json '{"query_text":"community repo","namespace":"default","allowed_buckets":["research"],"sensitivity_ceiling":"standard"}'Repository overview:
Dashboard preview:
.
├── src/neurocore/ # Core package, adapters, storage, retrieval, reporting
├── tests/ # Pytest suite grouped by subsystem
├── scripts/ # Bootstrap, checkout, proof, and repo helper scripts
├── assets/ # Banner and screenshots
├── docs/ # Setup, security, MCP, troubleshooting, and SSD docs
├── dashboards/ # Dashboard contribution surface and templates
├── extensions/ # Curated extension surface and bundle manifests
├── integrations/ # Connector and starter integration examples
├── primitives/ # Reusable building blocks and templates
├── recipes/ # Runnable workflow recipes
├── schemas/ # Checked-in contracts and schema templates
├── skills/ # Skill definitions and templates
├── .claude/commands/ # Repo-local slash-command prompts
└── .github/ # CI workflows, templates, and metadata schemas
- docs/README.md for the docs index
- docs/setup.md for the full bootstrap and manual setup flow
- docs/reference-stack.md for the default local path
- docs/hosted-stack.md for the hosted companion path
- docs/security-workflows.md for security-focused memory workflows
- docs/security.md for publication and runtime hygiene
- docs/mcp/CODEX.md and docs/mcp/CLAUDE_CODE.md for local MCP setup
Use the repo entrypoints instead of ad hoc commands:
make test
make lint
make validate
make openapi-checkDo not commit API keys, local .env files, database files, token.json,
secrets.json, or preferences.json.
- Copy
.env.exampleinto your operator-home env file, not into the repo root. - Treat
secrets.json.example,preferences.json.example, andingest-profiles.json.exampleas templates only. - Sanitize provider URLs, headers, and logs before opening issues or sharing test output.
- Use SECURITY.md for vulnerability reporting guidance.
Common setup and runtime issues are documented in docs/troubleshooting.md.
