=== Medusa Skill Audit Report ===

Skill: ai-ml (ai-ml), level: Godlike
  Experience: 100.0/100
  Confidence: 75%
  Metrics:
    - Content Length: 5966 chars
    - Code Blocks: 15
    - Step Instructions: 0
    - Technical Terms: 26
    - Complexity Score: 80.0/100
    - Value Score: 90.0/100

Running A/B Test: Parallel vs Sequential Scan
Path: /path/to/skills
Iterations: 20

Hypothesis: Parallel scanning is faster than sequential
Primary metric: scan_time_ms

Iteration 1: Parallel=178ms, Sequential=362ms
Iteration 2: Parallel=187ms, Sequential=325ms
...
Iteration 20: Parallel=204ms, Sequential=391ms

=== A/B Test Results ===
Parallel avg: 190.00ms
Sequential avg: 352.00ms
✅ Parallel is 46.0% faster

Medusa Skill Framework (MSF) v0.12 - Audit-Based Ranking with Context
Usage: medusa <command> [options]

Commands:
  scan <path>              Scan skills with FULL audit (60/30/10 scoring)
    --sequential           Use sequential scanning (no Rayon)
    --no-cache             Disable incremental scan cache

  audit <path>             Show detailed skill audit with cross-session context
    --no-cache             Disable cache

  html <path> <output>     Generate HTML visualization
    --sequential           Use sequential scanning
    --no-cache             Disable cache

  export-csv <path> <f>    Export skills to CSV format
  export-md <path> <f>     Export skills to Markdown
  export-svg <path> <f>    Export skills to SVG visualization

  ab-test <path>           Run A/B test (parallel vs sequential)
    --iterations N         Number of test iterations (default: 10)

  dream <path>             Run dreaming process (cross-session pattern detection)
  dream-status <path>      Show dream knowledge base and patterns
  dream-reset <path>       Reset dream state and history
  dream-consolidate <path> Manually consolidate dream knowledge base
  dream-diary <path>       Show dream diary (narrative skill evolution timeline)
    --output <file.md>     Export diary as Markdown
  dream-params <path>      Show dreaming configuration parameters

  orchestrate <path>       Run multi-agent orchestrated audit (4 specialized sub-audits)
    --sequential           Use sequential scanning
    --no-cache             Disable cache

  outcome-add <path> <id>  Add default outcome rubric for a skill
  outcome-list <path>      List outcome rubrics
  outcome-remove <path> <id> Remove an outcome rubric
  learning-path <path> <id> Show learning path and suggestions for a skill

  procedural-list <path>   List all learned procedural workflows
  procedural-show <p> <id> Show workflows associated with a skill

  memory-export <p> <f>    Export all memory (dream, procedural, outcomes) to JSON bundle
  memory-import <p> <f>    Import and merge a memory bundle from another Medusa instance
    --source <name>        Tag imported data with a source identifier

  update                  Update Medusa from GitHub (git pull + rebuild)

Options:
  --help, -h              Show this help message
  --version, -v           Show version

SKILL.md files
    ↓
[WalkDir] Scan filesystem (max depth 4)
    ↓
[Rayon] Parallel processing (46% faster, optional)
    ↓
[Regex] Extract YAML frontmatter
    ↓
[Audit] Measure complexity (length, code, steps, terms)
    ↓
[Score] Calculate experience (60% + 30% + 10%)
    ↓
[Rank] Assign tier (Godlike → Poor)
    ↓
[Fusion] Detect similar skills (FxHash)
    ↓
[Session] Record snapshot for dreaming
    ↓
[Dream] Cross-session pattern detection + consolidation
    ↓
[Diary] Generate skill evolution timeline
    ↓
[Agents] Multi-agent orchestrated sub-audits
    ↓
[Outcomes] Rubric-based quality assessment
    ↓
[Procedural] Extract step-by-step workflows
    ↓
[Output] JSON / HTML / CSV / MD / SVG

medusa/
├── src/
│   ├── main.rs          # CLI entry, scoring, reports (~1350 lines)
│   ├── dream.rs         # Dreaming, consolidation, diary, learning paths
│   ├── outcomes.rs      # Rubric CRUD, skill assessment
│   ├── agents.rs        # Multi-agent orchestrated audit (4 agents)
│   └── procedural.rs    # Procedural workflow detection & memory
├── target/
│   └── release/
│       └── medusa       # Compiled binary
├── Cargo.toml          # Dependencies (minimal: 9 deps)
├── medusa.toml         # Configurable scoring + dreaming params
├── README.md           # This file
├── TECHNICAL.md        # Architecture deep-dive
└── .medusa_state.json  # Promotion state (auto-created)

serde       — Struct serialization
serde_json  — JSON output
toml        — Config parsing
fxhash      — Fusion hash computation
walkdir     — Directory traversal
rayon       — Parallel processing
regex       — Pattern extraction
lazy_static — Regex compilation
chrono      — Timestamps for dream sessions

=== Medusa Skill Audit Report ===

Skill: ai-ml (ai-ml), level: Godlike
  Experience: 100.0/100
  Confidence: 75%
  Metrics:
    - Content Length: 5966 chars
    - Code Blocks: 15
    - Step Instructions: 0
    - Technical Terms: 26
    - Complexity Score: 80.0/100
    - Value Score: 90.0/100

Medusa Skill Framework (MSF) v0.12.1