I decided to run music as a documentation-first operating system instead of a code-heavy product: the repo is the control plane for strategy, releases, growth experiments, and weekly execution, with decisions captured in Markdown so future moves stay consistent.

What We Built

• A working knowledge base centered on planning and operating an artist project in files like music-crm.md, music-kpi-tracker.md, music-weekly-log.md, and music-history.md.
• A repeatable operating loop defined in README.md: capture research, promote stable facts to CRM, snapshot KPIs weekly, append weekly logs, and only record major milestones in history.
• A growth playbook surface for DJ-led distribution and testing (dj-growth-engine.md, trend-discovery-system.md, playlist and Spotify strategy notes).
• A memory layer (MEMORY.md) that locks in key constraints: this is year-zero from the first release (Renegade Bassline EP), and process decisions should favor practical, repeatable execution.

Why We Built It

• Recent sessions show the same pressure from multiple angles: improve Spotify growth, build a recurring SoundCloud mix lane, and run playlist submissions without losing focus.
• The right tradeoff here is speed and clarity over tooling complexity; there are no package scripts or automation requirements yet, so Markdown remains the fastest reliable interface.
• We need durable decision memory so strategy does not reset every week; this is especially important when testing formats, snippets, and release sequencing from scratch.
• The AGENTS guidance reinforces this: shared memory is canonical, and research work should be captured and reused rather than rebuilt each session.

How It Works

• Research and platform findings land first in music-platform-research.md; only verified, stable details get promoted into music-crm.md.
• Weekly execution runs through music-kpi-tracker.md and music-weekly-log.md, keeping performance and priorities visible over time.
• Milestones and durable timeline events are preserved in music-history.md so tactical experiments do not pollute long-term recordkeeping.
• Strategy files (for trend discovery, DJ clip systems, and channel growth experiments) act as reusable SOPs that can be iterated without changing the core operating model.
• Current state is documentation-led with no recent commit trail, so operational discipline depends on keeping these core files current each week.