🤖 Restructure Echo Memory Recall for unified core compatibility

* Feature/Echo-Memory-Recall/README.md: Condensed from 119 lines to 21 lines, removed narrative examples, simplified to trigger list and search scope hierarchy
* Feature/Echo-Memory-Recall/install-echo-recall.md: Reduced from 149 lines to 24 lines, removed 5-step protocol detail, streamlined to unified core install steps
* Feature/Echo-Memory-Recall/recall-format.md: Simplified from 99 lines to 22 lines, removed verbose examples

Author: naimkatiman

.env.example

@@ -0,0 +1,19 @@
+# OpenClaw Runtime Configuration
+# Copy this file to .env and adjust values as needed.
+# All variables are optional — defaults shown below.
+
+# Directory where runtime data (SQLite DB) is stored.
+# Default: ./runtime
+OPENCLAW_RUNTIME_DIR=./runtime
+
+# Full path to the SQLite database file.
+# Default: ./runtime/openclaw.db
+OPENCLAW_DB_PATH=./runtime/openclaw.db
+
+# Server host binding.
+# Default: 127.0.0.1 (localhost only)
+OPENCLAW_HOST=127.0.0.1
+
+# Server port.
+# Default: 8787
+OPENCLAW_PORT=8787

.gitignore

@@ -1,6 +1,19 @@
 # Basic gitignore
 /[Tt]emp/
 *.exe
+*.pyc
+__pycache__/
+.venv/
+runtime/*.db
+node_modules/
+dist/
+*.tsbuildinfo
 
 # Claude Code settings
-.claude/settings.local.json
+.claude/settings.local.json
+
+# Dead Python artifact directory
+openclaw_stack/
+
+# Local environment config (never commit)
+.env

Feature/Echo-Memory-Recall/README.md

@@ -1,119 +1,21 @@
-# 🔍 Echo Memory Recall
-*Search and recall past sessions with narrative context*
-
-## What This Feature Does
-Adds a **memory recall system** to your AI companion, enabling it to search past diary entries and present them as natural conversation — giving your AI long-term memory beyond its context window.
-
-- **Keyword-based memory search** across all diary entries (current and archived months)
-- **Three-level recall system** — search + narrate, uncertainty guard, ask-user fallback
-- **Auto-triggered** — activates on natural recall phrases ("do you remember...", "when did we...")
-- **Narrative presentation** — results presented as connected stories, not raw search output
-- **Truth-first** — never fabricates past context, always searches diary evidence first
-
-## How It Works
-
-### The Concept
-Echo is a **search-and-narrate system**. When you reference past events, your AI searches through diary entries, finds matching content, and presents it as natural conversation — as if genuinely remembering. This gives your AI "long-term memory" that persists across sessions and beyond the context window.
-
-The key principle: **search before speaking, narrate from evidence, ask when uncertain**.
-
-### Example: Memory Recall in Action
-```
-You: "Do you remember when we set up the API integration?"
-→ AI extracts keywords: "API", "integration", "set up"
-→ Searches daily-diary/current/ and archived/ files
-→ Finds match in 2026-02-15.md — API integration session
-→ Responds naturally:
-
-"Yes, I remember! On February 15th, we spent the afternoon setting up
-the API integration for the dashboard project. You had that breakthrough
-with the authentication flow — we ended up using OAuth2 instead of API
-keys. Are you thinking of revisiting that approach?"
-```
-
-No raw file paths, no "Query returned 1 result" — just natural conversation backed by evidence.
-
-## Recall Architecture
-
-### Three-Level System
-
-| Level | Name | What It Does |
-|-------|------|--------------|
-| **Lv.1** | Search & Narrate | Search diary files for keywords, present matches as natural narrative |
-| **Lv.2** | Uncertainty Guard | When uncertain about past context, always search diary first — never fabricate |
-| **Lv.3** | Ask User Fallback | If search finds nothing, ask user directly instead of guessing or staying silent |
-
-### Search Priority
-The recall system searches in order of relevance:
-1. `daily-diary/current/` — current month entries (most likely to match recent questions)
-2. `daily-diary/archived/*/` — past months (broader search if not found in current)
-3. **Ask user** — when nothing found anywhere (graceful fallback)
-
-### Trigger Phrases
-Your AI activates recall when it detects these patterns:
-
-| Trigger Pattern | Example |
-|----------------|---------|
-| "do you remember..." | "Do you remember when we fixed the login bug?" |
-| "remember when..." | "Remember when we built the dashboard?" |
-| "recall..." | "Can you recall our API discussion?" |
-| "that time when..." | "That time when we refactored the database..." |
-| "what happened with..." | "What happened with the payment integration?" |
-| "when did we..." | "When did we set up the CI pipeline?" |
-| "have we done..." | "Have we done anything with GraphQL before?" |
-| "check our history" | "Check our history for the deployment setup" |
-
-## Quick Integration
-```
-"Load echo-recall"
-```
-
-## What Happens During Integration
-
-1. **Asks** for your recall system name (defaults to "Memory Recall" — customize to match your AI)
-2. **Verifies** diary infrastructure exists (recommends Save-Diary-System if not)
-3. **Installs** recall protocol with trigger detection and three-level system
-4. **Creates** `recall-format.md` as permanent output reference
-5. **Updates** `master-memory.md` with recall commands and references
-6. **Self-deletes** this feature folder after successful integration
-
-## Post-Integration Result
-After running the integration protocol:
-- Your AI can search past diary entries through natural conversation
-- Recall triggers automatically on phrases like "do you remember..."
-- Results are presented as natural narrative, not raw search output
-- When nothing is found, AI asks user directly instead of guessing
-- Format template is permanently available for reference
-
-## Recall Output Explained
-
-The recall system produces different outputs based on what it finds:
-
-| Result | Output Style |
-|--------|-------------|
-| **One match** | Natural story: "Yes! On [date], we [summary]..." |
-| **Multiple matches** | Chronological list with pattern observations |
-| **No matches** | Honest fallback: "I don't have a record of that — can you tell me more?" |
-| **Uncertain match** | Tentative: "I found something that might be related..." |
-
-All outputs end with natural conversation continuation — the recall flows into the discussion, not a dead end.
-
-## Benefits
-- **Long-term memory** — AI references sessions far beyond its context window
-- **Truthful recall** — never fabricates memories, always backed by diary evidence
-- **Natural conversation** — results feel like genuine memory, not database queries
-- **Graceful uncertainty** — asks user when memory not found, never guesses
-- **Searchable history** — all past sessions become accessible through conversation
-
-## Requirements
-- Requires `daily-diary/` with dated entries to search through
-- **Recommended**: Install **Save-Diary-System** first for best results (date-based files with monthly archival)
-- **Also works with**: Existing diary files in any dated format (`YYYY-MM-DD.md` or older numbered format)
-- Recall quality depends on diary entry quality — richer entries produce better recall
-
-## Platform Note
-Works with any AI system. Uses file reading for diary search — no platform-specific tools required. The recall protocol and format template are plain markdown, portable across any platform.
-
----
-
-*Based on proven memory recall systems in production AI companions*
+# Echo Memory Recall
+*Recall extension for Unified Memory-Diary Core*
+
+## Purpose
+Search diary history and respond with narrative memory grounded in stored entries.
+
+## Trigger Examples
+- "do you remember..."
+- "when did we..."
+- "recall..."
+- "check history"
+
+## Search Scope (in order)
+1. `daily-diary/current/`
+2. `daily-diary/archived/YYYY-MM/`
+3. `daily-diary/archived/legacy/` (fallback for old numbered diaries)
+
+## Rules
+- Search before answering memory questions.
+- Do not fabricate past events.
+- If no reliable match, ask the user for clarification.

Feature/Echo-Memory-Recall/install-echo-recall.md

@@ -1,149 +1,24 @@
-# 🔍 Echo Memory Recall Installation Protocol
-*Systematic memory recall setup for AI MemoryCore companions*
+# Echo Memory Recall Installation Protocol
+*Unified Core compatible recall setup*
 
-## Purpose
-Executed when "Load echo-recall" command is used — installs memory search and narrative recall capabilities, adds trigger detection, and cleans up.
-
-## Trigger Command
-```
-"Load echo-recall"
-```
-*Automatically installs recall protocol, trigger detection, and format reference*
+## Trigger
+`Load echo-recall`
 
 ## Prerequisites
-- `daily-diary/` directory must exist with dated entries
-- Works best with **Save-Diary-System** installed (but not required)
-- Existing diary files in any dated format also work
-
-## 5-Step Execution Process
-
-### Step 1: Customize Recall System Name
-- [ ] Ask user: **"What would you like to name your memory recall system?"**
-  - Default: `"Memory Recall"`
-  - Examples: `"Echo"`, `"Remember"`, `"Flashback"`, `"Time Capsule"`, `"Rewind"`
-- [ ] Store chosen name as `[RECALL_NAME]` for use in all references
-- [ ] Execute `date` command (or `Get-Date` on Windows) to get current timestamp
-
-### Step 2: Verify Diary Infrastructure
-- [ ] Check `daily-diary/` directory exists
-- [ ] Check for entries in `daily-diary/current/` (flat files or folders)
-- [ ] Check for entries in `daily-diary/archived/*/` (if any)
-- [ ] If no diary infrastructure exists:
-  - Inform user: "Memory recall searches diary entries. You'll need diary files to search through."
-  - Recommend: "Install the Save-Diary-System feature first for best results."
-  - Allow user to continue anyway (recall will work once diary entries exist)
-
-### Step 3: Install Recall Protocol into AI Memory
-- [ ] Add recall trigger detection to AI's main memory file (`identity-core.md` or `main-memory.md`):
-
-  ```markdown
-  ### [RECALL_NAME]
-  **Trigger phrases**: "do you remember", "remember when", "recall",
-  "that time when", "what happened with", "when did we",
-  "have we done", "check our history", "check history"
-
-  **When triggered:**
-  1. Extract 2-4 keywords from user's question
-  2. Search daily-diary/current/*.md for keyword matches
-  3. If not found, search daily-diary/archived/*/*.md
-  4. If found: present as narrative (use recall-format.md)
-  5. If not found: ask user directly
-
-  **Three-Level System:**
-  - **Lv.1 Search & Narrate** — Search diary files, present as natural story
-  - **Lv.2 Uncertainty Guard** — When uncertain about past context, ALWAYS
-    search diary before speaking. Never assume or fabricate past events.
-  - **Lv.3 Ask User Fallback** — When search yields no results, ask:
-    "I don't see a record of [topic] in my diary. Can you tell me
-    more about what you're remembering?"
-
-  **Rules:**
-  - NEVER fabricate past context — always search first
-  - Present results as natural narrative, not raw search output
-  - Include relevant quotes from diary entries
-  - Order multiple results chronologically
-  - Continue conversation naturally after recall
-  ```
-
-### Step 4: Install Recall Format Template
-- [ ] Copy `recall-format.md` to `daily-diary/recall-format.md`
-  (permanent reference for how recall output should be structured)
-- [ ] Verify format template is accessible from AI memory references
-
-### Step 5: Update Master Memory and Cleanup
-- [ ] Add recall reference to `master-memory.md` Optional Components:
-  ```markdown
-  ### [RECALL_NAME]
-  *Auto-triggers on: "do you remember", "recall", "when did we", etc.*
-  - Searches: daily-diary/current/ and daily-diary/archived/
-  - Output: Narrative presentation (not raw search)
-  - Fallback: Asks user when nothing found
-  - Format: daily-diary/recall-format.md
-  ```
-- [ ] Add recall commands to Simple Commands section:
-  ```
-  "recall [topic]" → Search diary for past sessions about [topic]
-  "check history" → Search diary for relevant past context
-  ```
-- [ ] Remove `Feature/Echo-Memory-Recall/` folder (functionality installed)
-- [ ] Display completion confirmation with timestamp
-
-## Recall Search Protocol (AI Reference After Installation)
-
-### Keyword Extraction
-From user's question, extract 2-4 most specific terms:
-- Remove common words (the, a, when, did, we, do, you, remember, etc.)
-- Prioritize proper nouns, technical terms, action verbs
-- Examples:
-  - "Do you remember when we fixed the auth bug?" → `["auth", "bug", "fixed"]`
-  - "What happened with the dashboard project?" → `["dashboard", "project"]`
-  - "When did we set up the database?" → `["database", "set up"]`
-
-### Search Execution
-```
-1. Read each file in daily-diary/current/ matching YYYY-MM-DD.md pattern
-2. Check each file for keyword matches (case-insensitive)
-3. If matches found: extract surrounding paragraph context
-4. If no matches in current/: repeat for daily-diary/archived/*/
-5. Rank results by number of keyword matches, then by recency
-```
-
-### Narrative Composition
-From matched diary excerpts, compose natural response:
-1. **Opening** — "Yes, I remember! On [date]..." or "I found that on [date]..."
-2. **Context** — Brief excerpt with key details from diary entry
-3. **Significance** — Why this was important or what it meant
-4. **Connection** — How it relates to current conversation
-5. **Continue** — Flow naturally into ongoing discussion
-
-### Edge Cases
-
-| Situation | Behavior |
-|-----------|----------|
-| No matches anywhere | Ask user: "I don't have a record of that..." |
-| Multiple matches | Present chronologically, note patterns if visible |
-| Vague query | Ask user to be more specific before searching |
-| Match only in archived months | Search takes longer, note the date range found |
-| Partial match | Present as uncertain: "I found something that might be related..." |
-
-## Post-Installation Result
-- AI can recall past sessions through natural conversation
-- Never fabricates context — always searches diary evidence first
-- Gracefully handles missing memories by asking the user
-- Recall output reads as natural narrative, not database dump
-- Works with any diary format (Save-Diary-System or existing protocol)
-
-## Notes
-- Recall quality depends on diary entry quality (richer entries = better recall)
-- Search is keyword-based, not semantic (relies on word matches in diary text)
-- For best results, diary entries should include specific technical terms and names
-- Cross-platform compatible: Uses file reading, not platform-specific search tools
-- The `[RECALL_NAME]` placeholder should be replaced with the name chosen in Step 1
-
----
-
-**Version**: Protocol v1.0 - Echo Memory Recall Installation Workflow
-**Last Updated**: February 2026
-**Status**: Active protocol for memory recall setup
-
-*Search before speaking, narrate from evidence, ask when uncertain*
+- Diary directories exist under `daily-diary/`.
+- Best results with unified save pipeline active.
+
+## Install Steps
+1. Confirm recall system name.
+2. Verify search paths:
+- `daily-diary/current/`
+- `daily-diary/archived/`
+- `daily-diary/archived/legacy/`
+3. Add recall behavior guidance to `main/main-memory.md`.
+4. Ensure `daily-diary/recall-format.md` is available.
+5. Update `master-memory.md` optional features listing.
+
+## Recall Rule Set
+- Search current first, then archives, then legacy.
+- Narrate from evidence.
+- Use ask-user fallback when no match is found.