chore: bump v0.7.1 — Simplify MCP tools (53→42), SidMemo-only search, docs cleanup

Removed 11 niche/overlapping tools from MVP: macro_run, session_save/restore,
desk_health/conflicts/checkout, skill_create/list, okr_list/okr_update,
traceability_matrix. Consolidated context_packs + knowledge_context into
entity_context. Fixed entity-graph mode SidMemo regression. Updated all docs
to match current 42-tool surface.

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>

Author: Dchuong03

CHANGELOG.md

@@ -1,5 +1,27 @@
 # Changelog
 
+## [0.7.1] - 2026-03-12
+
+### Changed
+- **MCP Tools reduced 53 → 42:** Removed 11 niche/overlapping tools from MVP whitelist
+  - Removed: `macro_run`, `session_save/restore`, `desk_health/conflicts/checkout`, `skill_create/list`, `okr_list/okr_update`, `traceability_matrix`
+- **SidMemo full dependency:** `knowledge_search` now SidMemo-only (no keyword fallback)
+- **`entity_context` resilience:** Entity-graph mode no longer requires SidMemo (optional overlay)
+
+### Removed
+- **`context_packs` handler:** Deleted, replaced by `entity_context` RAG mode
+- **`memory_search` handler:** Removed dead code from memory.ts
+- **`knowledge_context` handler:** Logic consolidated into `entity_context`
+- **`quick_context` macro:** Removed from macros.ts
+- **Legacy tool references:** Cleaned up across docs, skills, agents, and templates
+
+### Fixed
+- `entity_context` SidMemo regression — entity-graph mode no longer fails without SidMemo
+- Tool count inconsistencies across documentation (now consistently 42)
+- Missing `knowledge_module_overview` from doc tables
+- Stale `session_launch` references in agent command templates
+- Duplicate `knowledge_search` line in review-workflow template
+
 ## [0.7.0] - 2026-03-06
 
 ### Added

CLAUDE.md

@@ -22,7 +22,7 @@ Keep this managed block so 'openspec update' can refresh the instructions.
 Task lifecycle, quality gates, lesson detection: managed by `sidstack-aware` skill. Every request is **classified by intent first** (via Workflow Router + sidstack-aware skill), then routed to the correct workflow. Use `/sidstack-dev` for explicit structured workflows. SidStack MCP tasks (`mcp__sidstack__task_*`) for persistence; built-in tasks for session-local sub-steps only.
 
 **Integrated workflow — always follow this pattern:**
-1. **Before work:** `knowledge_search` + `memory_search` → find context. `entity_link` → link relevant docs to task.
+1. **Before work:** `knowledge_search` → find context. `entity_link` → link relevant docs to task.
 2. **During work:** `entity_context` → get linked context. `task_update` → track progress.
 3. **After work:** `test_result_create` → persist test results. `memory_add` → store learnings. `task_complete` → finish.
 <!-- GOVERNANCE:END -->
@@ -137,7 +137,7 @@ pnpm typecheck            # Type checking
 
 ## Architecture Reference
 
-**MCP Tools:** knowledge (9), tasks (5), impact (3), tickets (4), training (8), OKRs (2), test results (3), agent desk (5), memory (6), traceability (1), entity references (3) — 49 tools.
+**MCP Tools:** knowledge (9), tasks (5), impact (3), tickets (4), training (6), test results (3), agent desk (4), memory (5), entity refs (3) — 42 tools.
 **Agents:** Worker (`sidstack-worker`) for implementation, Reviewer (`sidstack-reviewer`) for verification. Skills auto-trigger per role.
 **Knowledge:** `.sidstack/knowledge/` in 9 categories (`00-context` through `08-incidents`).
 **Impact:** `impact_analyze` → `impact_check_gate`. Gates: `blocked`, `warning`, `clear`.
@@ -159,7 +159,7 @@ pnpm typecheck            # Type checking
 | | `memory_add(content, projectId)` | Store learnings for future |
 | | `lesson_create(...)` | After fixing tricky bugs (ask user first) |
 
-> **Tip:** `entity_context` loads everything linked to a task (knowledge, memory, references) in one call — use it instead of calling `knowledge_search` + `memory_search` + `entity_references` separately.
+> **Tip:** `entity_context` loads everything linked to a task (knowledge, memory, references) in one call — use it instead of calling `knowledge_search` + `entity_references` separately. It also supports RAG mode with `projectPath` + `query`/`moduleId`.
 
 ---
 

README.md

@@ -17,8 +17,7 @@ Local-first. SQLite-based. No external services required.
 | **Impact Analysis** | Assess scope, risks, and blockers before making changes |
 | **Task Management** | Track AI work with governance and quality gates |
 | **Ticket Queue** | Intake external tickets, review, convert to tasks |
-| **Training Room** | Capture lessons from incidents, build reusable skills and rules |
-| **OKRs** | Define project goals, track progress through task completion |
+| **Training Room** | Capture lessons from incidents, build reusable rules |
 
 ## Quick Start
 
@@ -43,7 +42,7 @@ Or configure MCP manually:
 }
 ```
 
-Claude Code now has access to 53 MCP tools for knowledge, tasks, impact analysis, tickets, training, memory, macros, and session management.
+Claude Code now has access to MCP tools for knowledge, tasks, impact analysis, tickets, training, memory, and agent desk management.
 
 ### Desktop App (macOS)
 
@@ -59,12 +58,11 @@ pnpm install && pnpm packages:build && pnpm tauri:build
 
 See [Quick Start Guide](docs/QUICK_START.md) for full setup instructions.
 
-## MCP Tools (53)
+## MCP Tools
 
 ```
-# Knowledge (10) - understand your project
-knowledge_context        Build context for a task/module
-knowledge_search         Search across knowledge docs
+# Knowledge (9) - understand your project
+knowledge_search         Semantic search across knowledge and memory (SidMemo)
 knowledge_list           List available docs
 knowledge_get            Get single document with full content
 knowledge_modules        List modules with stats
@@ -92,60 +90,43 @@ ticket_list              List/filter tickets
 ticket_update            Update status
 ticket_convert_to_task   Convert to task
 
-# Training (8) - learn from mistakes
+# Training (6) - learn from mistakes
 incident_create          Report an incident
 incident_list            List incidents
 lesson_create            Create lesson from incident
 lesson_list              List lessons
-skill_create             Create reusable skill
-skill_list               List skills
 rule_check               Check rules for context
 training_context_get     Get training context for session
 
-# OKRs (2) - project goals
-okr_list                 List objectives and key results
-okr_update               Update key result progress
-
 # Test Results (3) - persist test execution
 test_result_create       Persist test execution results
 test_result_list         List test results with filtering
 test_result_get          Get detailed test result
 
-# Agent Desk (7) - workspace isolation
+# Agent Desk (4) - workspace isolation
 desk_create              Create agent desk (git worktree)
 desk_list                List all agent desks
 desk_status              Get desk status and current task
-desk_checkout            Switch desk to a task
-desk_health              Check desk health
-desk_conflicts           Detect merge conflicts
 desk_remove              Remove agent desk
 
-# Memory (6) - semantic search
+# Memory (5) - store learnings
 memory_add               Store a memory entry
-memory_search            Semantic search across memories
 memory_list              List memory entries
 memory_delete            Delete a memory entry
 memory_index_knowledge   Index knowledge docs into memory
 memory_cleanup           Clean up expired entries
 
-# Context & References (4)
+# Entity References (3)
 entity_link              Link entities (task↔knowledge, etc.)
 entity_references        Get references for an entity
-entity_context           Build context from linked entities
-traceability_matrix      Generate spec-task-test coverage matrix
-
-# Productivity (v0.7.0)
-context_pack             Build comprehensive context pack for a module
-macro_run                Run composite macros (start_work, finish_work, quick_context)
-session_save             Save session state for cross-session continuity
-session_restore          Restore saved session state
+entity_context           Build context from linked entities (entity mode + RAG mode)
 ```
 
 ## Desktop App - 7 Views
 
 | View | Shortcut | Description |
 |------|----------|-------------|
-| Project Hub | `Cmd+1` | Capability tree, entity connections, OKR progress |
+| Project Hub | `Cmd+1` | Capability tree, entity connections, project overview |
 | Task Manager | `Cmd+2` | List, kanban, timeline, detail views |
 | Knowledge Browser | `Cmd+3` | Tree view, search, type filtering |
 | Ticket Queue | `Cmd+4` | Status workflow, convert to tasks |
@@ -160,7 +141,7 @@ sidstack/
 ├── src/                   # React frontend (Tauri)
 ├── src-tauri/             # Rust backend (Tauri)
 ├── packages/
-│   ├── mcp-server/        # MCP Server (53 tools, npx-ready)
+│   ├── mcp-server/        # MCP Server (npx-ready)
 │   ├── api-server/        # REST API (Express.js, central gateway)
 │   ├── cli/               # CLI (init, update, scan, skills, desk)
 │   ├── bot-server/        # SidBot (Gemini intent router)

docs/CLAUDE_CODE_INTEGRATION.md

@@ -61,17 +61,17 @@ List my SidStack tasks
 
 Claude should use the `task_list` MCP tool.
 
-## Available MCP Tools (53)
+## Available MCP Tools
 
 ### Knowledge (9)
 
 | Tool | Description |
 |------|-------------|
-| `knowledge_context` | Build context for a task/module |
-| `knowledge_search` | Search across knowledge docs |
+| `knowledge_search` | Semantic search across knowledge and memory (SidMemo) |
 | `knowledge_list` | List available docs |
 | `knowledge_get` | Get single document with full content |
 | `knowledge_modules` | List modules with stats |
+| `knowledge_module_overview` | Get module architecture overview |
 | `knowledge_create` | Create knowledge document |
 | `knowledge_update` | Update knowledge document |
 | `knowledge_delete` | Delete knowledge document |
@@ -104,26 +104,17 @@ Claude should use the `task_list` MCP tool.
 | `ticket_update` | Update ticket status |
 | `ticket_convert_to_task` | Convert ticket to task |
 
-### Training (8)
+### Training (6)
 
 | Tool | Description |
 |------|-------------|
 | `incident_create` | Report an incident |
 | `incident_list` | List incidents |
 | `lesson_create` | Create lesson from incident |
 | `lesson_list` | List lessons |
-| `skill_create` | Create reusable skill |
-| `skill_list` | List skills |
 | `rule_check` | Check rules for context |
 | `training_context_get` | Get training context for session |
 
-### OKRs (2)
-
-| Tool | Description |
-|------|-------------|
-| `okr_list` | List objectives and key results |
-| `okr_update` | Update key result progress |
-
 ### Test Results (3)
 
 | Tool | Description |
@@ -132,51 +123,32 @@ Claude should use the `task_list` MCP tool.
 | `test_result_list` | List test results with filtering |
 | `test_result_get` | Get detailed test result |
 
-### Agent Desk (7)
+### Agent Desk (4)
 
 | Tool | Description |
 |------|-------------|
 | `desk_create` | Create agent desk (git worktree) |
 | `desk_list` | List all agent desks |
 | `desk_status` | Get desk status and current task |
-| `desk_checkout` | Switch desk to a task |
-| `desk_health` | Check desk health |
-| `desk_conflicts` | Detect merge conflicts |
 | `desk_remove` | Remove agent desk |
 
-### Memory (6)
+### Memory (5)
 
 | Tool | Description |
 |------|-------------|
-| `memory_add` | Add a memory entry (via mem0) |
-| `memory_search` | Semantic search across memories |
+| `memory_add` | Add a memory entry |
 | `memory_list` | List memory entries |
 | `memory_delete` | Delete a memory entry |
 | `memory_index_knowledge` | Index knowledge docs into memory |
 | `memory_cleanup` | Clean up stale memory entries |
 
-### Traceability (1)
-
-| Tool | Description |
-|------|-------------|
-| `traceability_matrix` | Generate spec-task-test coverage matrix |
-
 ### Entity References (3)
 
 | Tool | Description |
 |------|-------------|
 | `entity_link` | Link entities (task↔knowledge, etc.) |
 | `entity_references` | Get references for an entity |
-| `entity_context` | Build context from linked entities |
-
-### Productivity (4) — v0.7.0
-
-| Tool | Description |
-|------|-------------|
-| `context_pack` | Build comprehensive context pack for a module (knowledge + memory + tasks in one call) |
-| `macro_run` | Run composite macros: `start_work`, `finish_work`, `quick_context` |
-| `session_save` | Save session state (decisions, blockers, files, progress) for continuity |
-| `session_restore` | Restore saved session state from previous conversation |
+| `entity_context` | Build context from linked entities (entity mode + RAG mode) |
 
 ### Claude Code Hooks (Quality Gates)
 
@@ -216,7 +188,7 @@ pnpm test       # Must pass
    → task_create (auto-applies governance)
 
 2. "What knowledge do we have about the auth module?"
-   → knowledge_search + knowledge_context
+   → knowledge_search + entity_context (RAG mode)
 
 3. "Analyze the impact of this change"
    → impact_analyze
@@ -257,25 +229,11 @@ pnpm test       # Must pass
    → ticket_convert_to_task
 
 3. "Load context for this task"
-   → knowledge_context
+   → entity_context
 
 4. [Implement the fix]
 
 5. "Complete the task"
    → task_complete
 ```
 
-### OKR-Driven Work
-```
-1. "Show our project goals"
-   → okr_list
-
-2. "Create a task for KR-1.1"
-   → task_create
-
-3. [Implement the feature]
-
-4. "Complete the task and update OKR progress"
-   → task_complete
-   → okr_update
-```

docs/DEMO_SCENARIOS.md

@@ -71,8 +71,7 @@ Step-by-step scenarios showcasing SidStack's key features.
 
 1. **Open Project Hub** (`Cmd+1`)
 2. **View project stats** - module count, active tasks, knowledge documents
-3. **Check OKR progress** - objectives and key results
-4. **Click quick actions** - jump to any feature:
+3. **Click quick actions** - jump to any feature:
    - Task Manager - create and track tasks
    - Knowledge Browser - browse project documentation
    - Ticket Queue - manage external tickets
@@ -83,7 +82,6 @@ Step-by-step scenarios showcasing SidStack's key features.
 ### What to highlight
 - Central dashboard with unified view
 - Quick access to all features
-- OKR progress tracking
 
 ---
 

docs/GETTING_STARTED.md

@@ -26,7 +26,7 @@ Open Claude Code in your project. SidStack tools are available immediately:
 
 ```
 "Search knowledge about the auth module"
-→ Claude calls knowledge_search + knowledge_context
+→ Claude calls knowledge_search + entity_context (RAG mode)
 
 "Create a task to fix the login bug"
 → Claude calls task_create
@@ -53,11 +53,9 @@ SidStack is an **AI-Powered Project Intelligence Platform** for Claude Code.
 | **Task Management** | Track AI work with governance and quality gates |
 | **Impact Analysis** | Assess scope, risks, and blockers before making changes |
 | **Ticket Queue** | Intake external tickets, review, convert to tasks |
-| **Training Room** | Capture lessons from incidents, build reusable skills and rules |
-| **OKRs** | Define project goals, track progress through task completion |
+| **Training Room** | Capture lessons from incidents, build reusable rules |
 | **Agent Desk** | Isolated workspaces for AI agents via git worktrees |
-| **Memory** | Semantic search across project memories (via mem0) |
-| **Traceability** | Spec-task-test coverage matrix |
+| **Memory** | Semantic search across project memories (via SidMemo) |
 
 ---
 
@@ -159,13 +157,12 @@ pnpm install && pnpm packages:build && pnpm tauri:build
 
 | View | Description |
 |------|-------------|
-| **Project Hub** | Central dashboard with project stats, OKR progress, quick actions |
+| **Project Hub** | Central dashboard with project stats and quick actions |
 | **Task Manager** | Kanban, list, timeline views with governance |
 | **Knowledge Browser** | Browse and search project knowledge documents |
 | **Ticket Queue** | External ticket intake, review workflow, convert to tasks |
-| **Training Room** | Incidents, lessons, skills, and enforcement rules |
+| **Training Room** | Incidents, lessons, and enforcement rules |
 | **Agent Desk** | Manage isolated agent workspaces (git worktrees) |
-| **Traceability** | Spec-task-test coverage matrix |
 | **Docs** | Built-in documentation viewer |
 | **Settings** | Per-project configuration |