Remove outdated embedding/semantic search references

Search is BM25-only (via Tantivy). Local embedding models and the
cas models CLI were removed previously. Updated docs to reflect current state.

Author: assimelha

CLAUDE.md

@@ -54,8 +54,7 @@ crates/            # Workspace crates (cas-core, cas-store, cas-search, etc.)
 |-----------|---------|
 | `src/types/` | Core data: Entry, Task, Rule, Skill |
 | `src/store/` | Storage abstraction (SqliteStore primary) |
-| `src/search/` | BM25 + semantic hybrid search |
-| `src/embedding/` | Qwen3 embeddings via Candle |
+| `src/search/` | Full-text search (BM25 via Tantivy) |
 | `src/cli/` | Command handlers |
 | `src/hooks/` | Claude Code integration |
 
@@ -199,81 +198,6 @@ Patterns:
 - `Bash(cas task:*)` - Allow task operations only
 - `Bash(:* --help)` - Allow help for any command
 
-## Embedding Model Configuration
-
-CAS supports configurable embedding models for semantic search. Models can be local (downloaded) or API-based.
-
-### Available Models
-
-| Model | Dimension | Type | Provider |
-|-------|-----------|------|----------|
-| `Qwen/Qwen3-Embedding-0.6B` | 1024 | Local | Default |
-| `BAAI/bge-small-en-v1.5` | 384 | Local | Lightweight |
-| `BAAI/bge-base-en-v1.5` | 768 | Local | Balanced |
-| `sentence-transformers/all-MiniLM-L6-v2` | 384 | Local | Fast |
-| `text-embedding-3-small` | 1536 | API | OpenAI |
-| `text-embedding-3-large` | 3072 | API | OpenAI |
-| `voyage-3` | 1024 | API | Voyage |
-
-### Model Commands (Human CLI)
-
-```bash
-# List available models
-cas models list              # All models
-cas models list --local      # Local models only
-cas models list --api        # API models only
-cas models list --embedding  # Embedding models only
-cas models list --reranker   # Reranker models only
-
-# View current configuration
-cas models info
-
-# Set embedding model
-cas models set BAAI/bge-small-en-v1.5
-
-# Set reranker model
-cas models set Qwen/Qwen3-Reranker-0.6B --reranker
-
-# Check download status
-cas models status
-
-# Download models
-cas models download          # Embedding model
-cas models download --all    # Both embedding and reranker
-
-# Migrate embeddings after model change
-cas models migrate           # Preview
-cas models migrate --confirm # Execute
-```
-
-### Configuration
-
-Model settings are stored in `.cas/config.yaml`:
-
-```yaml
-embedding:
-  model: "Qwen/Qwen3-Embedding-0.6B"  # Embedding model ID
-  reranker: "Qwen/Qwen3-Reranker-0.6B"  # Optional reranker
-```
-
-### API Models
-
-For OpenAI or Voyage models, set environment variables:
-- `OPENAI_API_KEY` for OpenAI models
-- `VOYAGE_API_KEY` for Voyage models
-
-### Migration
-
-When changing to a model with different dimensions, existing embeddings must be regenerated:
-
-```bash
-# Set new model
-cas models set text-embedding-3-small
-
-# Migrate (regenerates all embeddings)
-cas models migrate --confirm
-```
-
 ## Code Search with ast-grep
 
 Use `ast-grep` for syntax-aware structural matching instead of text-only tools like `rg` or `grep`. This is especially useful for finding code patterns, refactoring, and understanding structure.

README.md

@@ -37,7 +37,7 @@ CAS is a local-first context system that gives AI agents persistent memory acros
 | **Tasks** | Track work with dependencies, priorities, and structured progress notes |
 | **Rules** | Coding conventions that earn trust through use and auto-sync to your editor |
 | **Skills** | Reusable agent capabilities with templates and usage tracking |
-| **Search** | Hybrid BM25 + semantic retrieval across all stored context |
+| **Search** | Fast full-text search (BM25) across all stored context |
 | **Factory** | Multi-agent orchestration — supervisor + workers in parallel git worktrees |
 
 ## Quick Start
@@ -198,7 +198,7 @@ CAS stores all data locally in your project:
 | `cas-cli` | CLI binary and MCP server |
 | `cas-core` | Core logic, hooks, and integrations |
 | `cas-store` | SQLite storage layer |
-| `cas-search` | BM25 + semantic hybrid search |
+| `cas-search` | Full-text search (BM25 via Tantivy) |
 | `cas-mcp` | MCP protocol handlers |
 | `cas-types` | Shared data types |
 | `cas-factory` | Multi-agent orchestration TUI |
@@ -210,26 +210,14 @@ CAS stores all data locally in your project:
 - **Rust** for performance and reliability
 - **SQLite** for local-first storage
 - **Tantivy** for full-text search (BM25)
-- **Qwen3 embeddings** for semantic search (runs locally)
 - **rmcp** for MCP protocol support
 - **Ratatui** for the factory TUI
 
 ## Configuration
 
-```yaml
-# .cas/config.yaml
-embedding:
-  model: "Qwen/Qwen3-Embedding-0.6B"    # Local embedding model
-  reranker: "Qwen/Qwen3-Reranker-0.6B"   # Optional reranker
-```
+CAS is configured via `.cas/config.yaml` in your project root. Run `cas config list` to see all options or `cas config describe <key>` for details on any setting.
 
 ```bash
-# List available embedding models
-cas models list
-
-# Switch models
-cas models set BAAI/bge-small-en-v1.5
-
 # Self-update
 cas update
 ```