A Claude Code configuration that gives your AI assistant a structured productivity layer. Goals, tasks, projects, pipeline evaluation, and compounding knowledge — all in plain markdown.
Built on Andrej Karpathy's insight that large language models are best understood not as chatbots, but as the kernel of a new kind of operating system.
"Think about it more like an operating system."
— Andrej Karpathy, Intro to Large Language Models (2023)
Three architectural views — system flow, project pipeline, and the compounding loop that makes each session smarter than the last.
From raw ideas to daily execution — everything flows through a single inbox, gets classified, and ties back to your goals.
/morning picks Top 5 daily/launch runs 7-stage pipelineEach project passes through seven stages — five evaluation stages with a Go/No-Go gate, a non-blocking technical-spec stage, and a final decomposition stage that promotes to active. Skip ahead with --from.
Three nested feedback loops. Each layer feeds the next — the system gets smarter over time.
/morning saves plans to journals. Next morning reads yesterday's actuals. Memories persist across sessions.
/weekly compiles shipping summaries, reads journals for plan-vs-actual patterns, proposes workflow improvements.
/quarterly scores OKRs, archives stale projects, refreshes GOALS.md, cleans stale memories, audits AGENTS.md.
A complete productivity system — from ideation to execution to compounding knowledge.
Clone the repo and run setup. Two paths — guided or manual.
setup.sh installs MCP server dependencies, creates workspace directories, and walks you through an interactive goals setup. Then run /refresh-goals in Claude Code to populate your goals.
| Tool | Version | Required | Install |
|---|---|---|---|
| Python | 3.11+ | Yes | brew install python@3.13 |
| uv | latest | Yes | curl -LsSf https://astral.sh/uv/install.sh | sh |
| git | any | Yes | brew install git |
| Claude Code | latest | Yes | claude.ai/download |
Each skill teaches your AI assistant a focused capability — from market validation to sprint planning.
Workflow automation at your fingertips — from daily standups to quarterly reviews.
--from.Each runs in its own context window, autonomously performing focused tasks in the background.
The manager-ai MCP server provides programmatic access to tasks, projects, and system state — with fuzzy deduplication built in.
| Tool | Description |
|---|---|
| list_tasks | Query tasks with filters (priority, status, category) |
| get_task_summary | Priority/category/status counts with time estimates |
| check_priority_limits | Alerts if P0 > 3 or P1 > 7 |
| prune_completed_tasks | Archive done tasks older than 30 days to tasks/archive/ |
| list_projects | Query projects with filters (status, priority, category) |
| get_pipeline_status | Count of projects at each pipeline stage |
| get_project_artifacts | Check which evaluation artifacts exist for a project |
| get_project_summary | Aggregate project stats and artifact coverage |
| get_system_status | Full dashboard — tasks + projects + backlog + time insights |
| process_backlog_with_dedup | Deduplicate backlog items against existing work |
The workflow is simple — brain dump, process, evaluate, execute, compound.
Connect external services to supercharge your workflows. All are optional — the core system works standalone.
/validate-project, /research-topic, and /discover-ideas. Uses the official Perplexity MCP server (@perplexity-ai/mcp-server).claude mcp add perplexity --env PERPLEXITY_API_KEY="your_key_here" -- npx -y @perplexity-ai/mcp-server
-s user before --env to register at user scope so every Claude Code project picks it up./morning and /weekly./plugin install slack
/meeting-sync. Requires paid plan..mcp.json. Authenticates via OAuth on first use. Setup guide/meeting-prep with email history and checking your schedule.brew install googleworkspace-cli
gws auth setup once to configure Google Cloud OAuth (requires gcloud). See README for full auth steps.Run one command after any change to skills, agents, hooks, or MCP tools — 37 deterministic checks catch drift before it ships.
Inline # /// script metadata auto-installs pyyaml. No venv setup.
CLAUDE.md, and every entry in CLAUDE.md exists on disk.server.py registrations, each tool has a wired dispatch branch.AGENTS.md tree matches reality; init-workspace.sh scaffolds the documented paths.evaluating/ready/active flagged when expected artifacts (validation, pre-mortem, PRD, stories) are missing.npm, gh, or gws warn if the CLI is not on PATH..DS_Store, stray TODO/FIXME markers, stale lock files, hardcoded user paths in the validator itself.Exit 0 clean, 1 on findings. Warnings (non-blocking) are reported separately. Run it before every PR.
./setup.sh. It installs MCP server dependencies, creates the workspace (tasks/, projects/, knowledge/) plus a blank BACKLOG.md and a GOALS.md template, then walks you through an interactive goals setup. Then run /refresh-goals in Claude Code for a guided walkthrough that fills in your goals, and /morning to start your first standup..claude/skills/<name>/SKILL.md and agents in .claude/agents/<name>.md. Each uses YAML frontmatter for configuration. Everything that you invoke as a slash command is also a skill — there's one consistent pattern. You can modify existing ones or create new ones by following the same shape.uv run core/scripts/validate.py. The validator runs 37 deterministic checks covering frontmatter, cross-references, registry parity, MCP tool wiring, workspace shape, pipeline conformance, external CLI deps, and hygiene. Exit 0 means clean. Warnings (non-blocking) are reported separately. Run it before every PR, or any time something feels off.