Files
foxhunt/.claude/commands/claude-flow-memory.md
jgrusewski 12fdd18223 refactor: remove entire CPU training path — 5,307 lines of dead code
Deleted:
- DQN::compute_loss_internal (280 lines) — old Candle forward+loss
- DQN::train_step (55 lines) — old Candle training step
- DQN::compute_gradients (47 lines) — old gradient accumulation
- ComputeLossResult struct — only used by deleted functions
- RegimeConditionalDQN::train_step (65 lines) — old dispatch
- RegimeConditionalDQN::train_step_gpu_regime (100 lines) — old GPU path
- RegimeConditionalDQN::compute_gradients_gpu (130 lines) — old regime gradients
- RegimeConditionalDQN::compute_gradients (92 lines) — old dispatch
- DQNAgentType::train_step dispatch — dead
- DQNAgentType::compute_gradients dispatch — dead
- GpuDqnTrainer::upload_batch (71 lines) — old CPU→GPU upload
- train_step.rs (500 lines) — entire module including ensure_fused_ctx
- dqn_benchmark.rs — used old train_step
- examples.rs — used old train_step
- validation/adapters.rs (289 lines) — used old train_step
- dqn/trainable_adapter.rs — used old train_step
- gpu_smoketest.rs — tested old train_step
- Gradient accumulation path in training_loop.rs (144 lines)
- IQN d_h_s2().clone() → raw pointer (zero alloc)
- Causal intervention format! string alloc removed
- Dead HER relabel functions (320 lines)

Kept:
- ensure_fused_ctx logic inlined into training_loop.rs
- set_noise_sigma_scale re-added to RegimeConditionalDQN

Fixed:
- GpuReplayBuffer max_batch_size wired from batch_size parameter
  (was hardcoded 1024, blocking batch_size=8192)

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
2026-04-02 09:06:23 +02:00

108 lines
2.8 KiB
Markdown

---
name: claude-flow-memory
description: Interact with Claude-Flow memory system
---
# 🧠 Claude-Flow Memory System
The memory system provides persistent storage for cross-session and cross-agent collaboration with CRDT-based conflict resolution.
## Store Information
```bash
# Store with default namespace
./claude-flow memory store "key" "value"
# Store with specific namespace
./claude-flow memory store "architecture_decisions" "microservices with API gateway" --namespace arch
```
## Query Memory
```bash
# Search across all namespaces
./claude-flow memory query "authentication"
# Search with filters
./claude-flow memory query "API design" --namespace arch --limit 10
```
## Memory Statistics
```bash
# Show overall statistics
./claude-flow memory stats
# Show namespace-specific stats
./claude-flow memory stats --namespace project
```
## Export/Import
```bash
# Export all memory
./claude-flow memory export full-backup.json
# Export specific namespace
./claude-flow memory export project-backup.json --namespace project
# Import memory
./claude-flow memory import backup.json
```
## Cleanup Operations
```bash
# Clean entries older than 30 days
./claude-flow memory cleanup --days 30
# Clean specific namespace
./claude-flow memory cleanup --namespace temp --days 7
```
## 🗂️ Namespaces
- **default** - General storage
- **agents** - Agent-specific data and state
- **tasks** - Task information and results
- **sessions** - Session history and context
- **swarm** - Swarm coordination and objectives
- **project** - Project-specific context
- **spec** - Requirements and specifications
- **arch** - Architecture decisions
- **impl** - Implementation notes
- **test** - Test results and coverage
- **debug** - Debug logs and fixes
## 🎯 Best Practices
### Naming Conventions
- Use descriptive, searchable keys
- Include timestamp for time-sensitive data
- Prefix with component name for clarity
### Organization
- Use namespaces to categorize data
- Store related data together
- Keep values concise but complete
### Maintenance
- Regular backups with export
- Clean old data periodically
- Monitor storage statistics
- Compress large values
## Examples
### Store SPARC context:
```bash
./claude-flow memory store "spec_auth_requirements" "OAuth2 + JWT with refresh tokens" --namespace spec
./claude-flow memory store "arch_api_design" "RESTful microservices with GraphQL gateway" --namespace arch
./claude-flow memory store "test_coverage_auth" "95% coverage, all tests passing" --namespace test
```
### Query project decisions:
```bash
./claude-flow memory query "authentication" --namespace arch --limit 5
./claude-flow memory query "test results" --namespace test
```
### Backup project memory:
```bash
./claude-flow memory export project-$(date +%Y%m%d).json --namespace project
```