wifi-densepose/vendor/ruvector/.claude/intelligence/IMPROVEMENTS.md

Intelligence System Improvements

Current State

• 5 hook types, 16 CLI commands
• 4,023 memories, 117 Q-states, 8,520 calibration samples
• Learning: command-type + context (cargo_in_rvlite, etc.)

Proposed Improvements

1. Error Pattern Learning (High Impact)

Learn from specific error types, not just success/failure.

// Instead of just: learn(state, 'command-failed', -0.5)
// Learn specific error patterns:
learn('cargo_build_error:E0308', 'type-mismatch', -0.3)
learn('cargo_build_error:E0433', 'missing-import', -0.2)

Benefit: Suggest fixes based on error type

2. File Sequence Learning (High Impact)

Track which files are often edited together.

// After editing lib.rs, user often edits:
sequences['crates/core/lib.rs'] = [
  { file: 'crates/core/tests/lib.rs', probability: 0.8 },
  { file: 'crates/core/Cargo.toml', probability: 0.3 }
]

Benefit: Proactively suggest related files

3. Crate Dependency Graph

Use the 42-crate structure for smarter suggestions.

dependencies = {
  'rvlite': ['ruvector-core', 'ruvector-attention-wasm'],
  'sona': ['ruvector-core']
}
// If editing rvlite, warn about downstream effects

Benefit: Warn about breaking changes

4. Test Suggestion Triggers

Automatically suggest running tests after certain edits.

// Post-edit hook detects:
if (file.match(/src\/.*\.rs$/) && !file.includes('test')) {
  suggest('Run tests: cargo test -p ' + crate);
}

Benefit: Reduce test-related bugs

5. Build Optimization

Learn minimal rebuild commands.

// Instead of 'cargo build', suggest:
if (changedCrates.length === 1) {
  suggest(`cargo build -p ${changedCrates[0]}`);
}

Benefit: Faster iteration cycles

6. Session Context Memory

Track patterns within the current session.

sessionContext = {
  filesEdited: ['lib.rs', 'mod.rs'],
  commandsRun: ['cargo check', 'cargo test'],
  errors: ['E0308 in line 45']
}
// Use for smarter in-session suggestions

Benefit: Context-aware suggestions

7. Git Branch Awareness

Different patterns for different branches.

// On feature branch: suggest more tests
// On main: suggest careful review
branchPatterns = {
  'main': { requireTests: true, suggestReview: true },
  'feature/*': { suggestTests: true }
}

Benefit: Branch-appropriate workflows

8. Hook Performance Metrics

Track hook execution time.

hookMetrics = {
  'pre-edit': { avgMs: 45, p99Ms: 120 },
  'post-command': { avgMs: 80, p99Ms: 200 }
}
// Alert if hooks become slow

Benefit: Prevent hook slowdowns

9. Predictive Prefetching

Pre-load likely-needed data.

// When user opens a Rust file, prefetch:
// - Related test files
// - Crate's Cargo.toml
// - Recent memories for that crate

Benefit: Faster responses

10. Multi-Crate Coordination

Optimize cross-crate work patterns.

// Detect multi-crate changes
if (editedCrates.length > 1) {
  suggest('Consider running: cargo build --workspace');
  recordPattern('multi-crate-edit', editedCrates);
}

Benefit: Better monorepo workflows

Implementation Priority

Improvement	Impact	Effort	Priority
Error Pattern Learning	High	Medium	1
File Sequence Learning	High	Medium	2
Test Suggestion	High	Low	3
Session Context	Medium	Medium	4
Build Optimization	Medium	Low	5
Crate Dependencies	Medium	Medium	6
Git Branch Awareness	Medium	Low	7
Hook Performance	Low	Low	8
Predictive Prefetch	Low	High	9
Multi-Crate Coord	Low	Medium	10