musk-algo

Musk Algo — Repo Flow Analyzer

Analyze the repo at $ARGUMENTS through two lenses:

1. Musk's 5-Step Algorithm — find what shouldn't exist at all
2. Theory of Constraints (Goldratt) — find what's slowing the entire system

Produce a deep, specific, actionable report with before/after examples and a prioritized roadmap. If $ARGUMENTS is empty, ask the user which repo path to analyze.

---

Phase 0: Repo Intake

Build a map of the repo before diving in:

# Structure
find <repo> -type f | grep -v node_modules | grep -v .git | grep -v dist | grep -v __pycache__ | head -100

# LOC by file
find <repo> -type f | grep -v node_modules | grep -v .git | xargs wc -l 2>/dev/null | sort -rn | head -30

# Dependency files
find <repo> -name "package.json" -o -name "requirements.txt" -o -name "Cargo.toml" -o -name "pom.xml" -o -name "go.mod" | grep -v node_modules

# Git depth
cd <repo> && git log --oneline | wc -l 2>/dev/null

# Largest files
find <repo> -type f -not -path "*/node_modules/*" -not -path "*/.git/*" | xargs ls -la 2>/dev/null | sort -k5 -rn | head -20

Build a mental model: project type, languages, scale (LOC, files, deps), activity (commits).

Detect repo type before choosing the scan depth

Some repos are 95% code. Others (Claude Code plugins, prompt libraries, documentation sites) are 95% markdown with a small code sidecar. The Musk/Goldratt lens still applies to prompt-heavy repos, but the scans in Phase 1 need to shift:

# What's the dominant file type?
find <repo> -type f -not -path "*/node_modules/*" -not -path "*/.git/*" -not -path "*/dist/*" \
  | sed 's/.*\.//' | sort | uniq -c | sort -rn | head -10

If .md dominates, treat prompts/skills as the "code" — scan them for stale references, half-finished renames, duplicated instructions, and sections that no longer match the current architecture. Tell the user explicitly: "This is a prompt-heavy repo, I'm shifting to a content-audit lens."

Tracked-artifact scan (things that shouldn't be in git at all)