Manager OpenBrain Starter

A local-first AI operating system for managers who are drowning in meetings, decisions, follow-ups, and scattered context.

Manager OpenBrain helps managers, product leads, program leads, and directors turn daily chaos into a repeatable operating rhythm. It gives them structured workflows for catch-ups, meeting memory, decisions, stakeholder briefs, risk reviews, and weekly program updates, while keeping sensitive work local-first and approval-gated.

Not a chatbot. Not a notes app. Not another dashboard.

This is a practical starter kit for making AI useful in the daily work of management without rebuilding the whole foundation from scratch.

This repo is intentionally generic. It should not contain company-specific systems, credentials, private documents, operational payloads, or live agent configuration.

Why Managers Would Want This

Managers do not want another tool to maintain. They want less rework, fewer forgotten decisions, and faster context recovery.

This starter gives them:

• A morning startup routine: what changed, what matters, and what should happen next.
• A restart routine for laptop restarts, context loss, vacations, or long gaps.
• Meeting memory so decisions, actions, open questions, and follow-ups do not disappear.
• A weekly program rhythm for risks, dependencies, movement, decisions, and stale assumptions.
• Safer AI use: read-only first, draft-only writes, human approval before posting or storing memory.
• Optional local memory with Qdrant and Docker for durable recall without sending everything to a cloud memory service.
• A reusable scaffold: prompts, workflows, schemas, examples, scripts, skills, and guardrails in one place.

What This Is

Manager OpenBrain is a local-first scaffold for:

• daily catch-up and priority triage
• meeting notes, decisions, and action tracking
• weekly program review
• stakeholder briefs
• risk and dependency review
• connector health checks
• tool-surface audits
• memory governance
• optional local Qdrant memory setup
• retrieval quality feedback

Principles

• Read-only first.
• Draft-only writes.
• Human approval before external writes.
• Human approval before memory writes.
• Store source pointers and approved summaries, not raw dumps.
• Treat embeddings and retrieval logs as sensitive.
• Keep secrets and live client config out of Git.
• Prefer local memory until an always-on need is proven and approved.

Trust And Safety

This repo is designed to be safe to fork and adapt:

• CI checks repo hygiene, broken links, malformed JSON, workflow manifest completeness, executable scripts, and secret-shaped content.
• Branch protection requires the scaffold validation workflow before changes land on main.
• The repo includes a security policy, contribution guide, code of conduct, support guide, changelog, issue templates, and pull request template.
• Local memory is optional and binds Qdrant to 127.0.0.1 by default.
• No live connectors, credentials, raw exports, or company-specific configuration are included.

Repo Shape

docs/
workflows/
workflow-packs/
prompts/
schemas/
examples/
skills/
scripts/
templates/

Start Here

1. Read docs/quickstart.md.

2. Read docs/operating-principles.md.

3. Review SECURITY.md and CONTRIBUTING.md before sharing changes.

4. Review CHANGELOG.md when you want to understand recent scaffold changes.

5. Start or restart the day with the local-only morning cockpit:

   scripts/openbrain_morning_start.py

   See docs/daily-operating-regime.md for the daily loop and day-by-day cadence.

6. Run the repo hygiene check:

   scripts/validate_repo_hygiene.py

7. Optional: if you want local vector memory, follow docs/local-memory-setup.md.

8. Start with these workflow packs:

   • workflows/mcp-tool-surface-audit.md
   • workflows/connector-health-check.md
   • workflows/daily-catch-up.md
   • workflows/meeting-memory-tracker.md
   • workflows/weekly-program-review.md

What Not To Put Here

• .env files
• OAuth tokens
• API keys
• Claude/Codex live config
• raw email exports
• raw Jira/Linear ticket dumps
• raw Slack/Teams exports
• confidential customer data
• meeting transcripts unless explicitly approved and sanitized

Current Status

Usable starter scaffold. It includes workflow packs, prompts, schemas, examples, repo hygiene CI, contribution guidelines, security policy, optional local Qdrant setup, and a morning/restart routine. No live connectors are configured by this repo. No memory backend is required to use the workflow templates.

Project Governance

• LICENSE: MIT.
• SECURITY.md: vulnerability reporting and security boundaries.
• CODE_OF_CONDUCT.md: expected behavior for project spaces.
• SUPPORT.md: where to get help and what is out of scope.
• CONTRIBUTING.md: contribution scope and safety checklist.