feat: Consolidate memory skills & punch up README

Memory Skill:
- Merged remember-this into unified memory skill
- Now uses Claude OS KBs (not separate memory MCP)
- Supports natural language: "remember this:", "save to memory:", etc.
- Deleted redundant remember-this skill

README:
- New tagline: "Give Your AI a Memory"
- Added "The Magic" section showing natural language UX
- Rewrote "What is Claude OS" with before/after comparison
- Added killer line: "Claude is a tool → Claude is a partner"
- Compact features table replacing wall of checkmarks

🤖 Generated with [Claude Code](https://claude.com/claude-code)

Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
This commit is contained in:
Bob Roberts 2026-01-02 16:04:31 -06:00
parent b3f9123c5b
commit 90de2e32e4
17 changed files with 253 additions and 340 deletions

View file

@ -0,0 +1,65 @@
{
"permissions": {
"allow": [
"SlashCommand(/statusline:preview-statusline)",
"Bash(test:*)",
"Bash(exit 0)",
"Bash(chmod:*)",
"Read(//Users/iamanmp/.claude/**)",
"Bash(pip3 list:*)",
"Bash(pip3 install:*)",
"Bash(pytest:*)",
"Bash(python3.9:*)",
"Bash(python3 -m pytest:*)",
"Bash(python3:*)",
"Bash(tee:*)",
"Read(//private/tmp/**)",
"Bash(git add:*)",
"Bash(git commit:*)",
"Bash(git log:*)",
"Bash(git rev-parse:*)",
"Bash(git push:*)",
"Read(///**)",
"Bash(git rev-list:*)",
"Bash(gh pr create:*)",
"Bash(gh issue edit:*)",
"Bash(gh pr view:*)",
"Read(//Users/iamanmp/**)",
"Bash(curl:*)",
"Bash(sqlite3:*)",
"Bash(ls:*)",
"Bash(cat:*)",
"Bash(claude --version)",
"Bash(find:*)",
"Bash(python test_session_implementation.py:*)",
"Bash(source venv/bin/activate)",
"Bash(pgrep:*)",
"Bash(ps:*)",
"Bash(npm run build:*)",
"Bash(npx tsc:*)",
"WebSearch",
"WebFetch(domain:github.com)",
"WebFetch(domain:skillsmp.com)",
"WebFetch(domain:docs.anthropic.com)",
"WebFetch(domain:docs.claude.com)",
"Bash(pkill:*)",
"Bash(git check-ignore:*)",
"Bash(/Users/iamanmp/Projects/claude-os/venv/bin/python -m pytest:*)",
"Bash(pip show:*)",
"Bash(pip install:*)",
"Skill(memory)",
"WebFetch(domain:thebob.dev)",
"Bash(git -C /Users/iamanmp/Projects/claude-os branch:*)",
"Bash(npx playwright:*)",
"Bash(source:*)",
"WebFetch(domain:localhost)",
"Bash(./stop_all_services.sh:*)",
"Bash(./start_all_services.sh:*)",
"Bash(grep:*)",
"mcp__code-forge__search_knowledge_base",
"mcp__code-forge__upload_document"
],
"deny": [],
"ask": []
}
}

View file

@ -118,9 +118,8 @@ These commands are installed to `~/.claude/commands/` via the install script:
| Skill | Purpose |
|-------|---------|
| `memory` | Save and recall information |
| `memory` | Save and recall information (supports trigger phrases like "remember this:") |
| `initialize-project` | Analyze codebase and generate docs |
| `remember-this` | Auto-save on trigger phrases |
---

View file

@ -8,8 +8,8 @@
</p>
<p align="center">
<strong>AI Memory & Knowledge Base System for Claude Code</strong><br>
Initialize projects in seconds • Never lose context again
<strong>Give Your AI a Memory</strong><br>
<em>Claude Code that actually remembers you.</em>
</p>
<p align="center">
@ -42,6 +42,24 @@
---
## ✨ The Magic
**You:** "Remember this: our API uses JWT tokens with 15-minute expiry and 7-day refresh tokens"
**Claude:** *Saved: "API Authentication Strategy" (Architecture)*
That's it. Next week, next month, next project - Claude remembers. No commands to memorize. No complex setup. Just talk naturally.
```
"remember this..." → saved to your knowledge base
"what did we decide?" → searches your memories
"how did we fix that?" → finds past solutions
```
**Every conversation makes Claude smarter. Every insight builds your shared knowledge.**
---
## 🆕 What's New in v2.3
> **Latest Release: December 2025**
@ -123,35 +141,55 @@ Thanks to our amazing contributors!
## 🚀 What is Claude OS?
**Claude OS** is **Claude Code's personal memory system** - making AI the best coding assistant in the universe by remembering everything across sessions.
Claude OS isn't just a tool - **it's Claude's memory**.
### The Problem
Think about it: you and Claude work together on a feature. You explain your architecture, your patterns, your preferences. Then you close the terminal... and it's all gone. Tomorrow, you start over.
You work with Claude Code on a feature, close the terminal, come back tomorrow... and Claude forgot everything. You explain the same architecture. You reference the same files. You repeat yourself constantly.
**What if Claude actually remembered?**
### The Solution
### Before Claude OS
**Claude OS gives Claude persistent memory:**
```
Day 1: "We use JWT tokens with refresh..."
Day 2: "As I mentioned, we use JWT tokens..."
Day 3: "Again, the auth system uses JWT..."
```
- 📝 **Remembers decisions** across all sessions
- 🔍 **Searches past work** automatically at session start
- 📚 **Indexes your docs** and makes them searchable
- 🧠 **Learns patterns** that improve over time
- 🔄 **100% Local** - Never leaves your machine, fully private
### After Claude OS
### Key Features
```
Day 1: "Remember this: we use JWT tokens with refresh..."
Day 2: Claude already knows. Applies the pattern automatically.
Day 3: Claude suggests improvements based on what worked.
```
✅ **Skills Library** - NEW! Browse & install skills from Anthropic and community repos
✅ **Session Insights** - NEW! Parse Claude Code sessions and extract patterns automatically
✅ **Lightning-Fast Indexing** - Tree-sitter hybrid indexing: 10,000 files in 30 seconds
✅ **Real-Time Kanban Board** - Auto-syncing task visualization for agent-os specs
✅ **One-Command Project Init** - `/claude-os-init` and you're done
✅ **Automatic Context Loading** - Starts every session with relevant memories
✅ **Session Management** - Track work, save progress, resume later
✅ **Documentation Ingestion** - Auto-indexes your docs during setup
✅ **Agent-OS Integration** - Optional spec-driven development with 8 specialized agents
✅ **Flexible AI** - Choose local Ollama (free) or OpenAI API (paid)
✅ **Template System** - Commands and skills shared via symlinks
### The Difference
| Without Claude OS | With Claude OS |
|-------------------|----------------|
| Explain the same things repeatedly | Claude remembers your decisions |
| Start cold every session | Context loaded automatically |
| Patterns forgotten | Patterns compound over time |
| Claude is a tool | Claude is a partner |
### Why It Works
- 🧠 **Persistent Knowledge** - Decisions, patterns, solutions saved across sessions
- 🔍 **Automatic Recall** - Relevant memories surface when you need them
- 📚 **Documentation Search** - Your docs indexed and searchable via RAG
- 🎯 **Session Learning** - Claude extracts insights from past conversations
- 🔒 **100% Local** - Your knowledge never leaves your machine
### Features at a Glance
| Feature | What It Does |
|---------|--------------|
| **Natural Language Memory** | Just say "remember this" - no commands needed |
| **Session Insights** | Auto-extracts patterns from past conversations |
| **Lightning Indexing** | 10,000 files indexed in 30 seconds |
| **Skills Library** | 36+ community skills, one-click install |
| **Cross-Project Learning** | Patterns from Project A help in Project B |
| **One-Command Setup** | `/claude-os-init` and you're ready |
---
@ -464,8 +502,7 @@ All these work in any initialized project:
**Global Skills (always available):**
- **`initialize-project`** - Analyze codebase and generate standards
- **`remember-this`** - Auto-save when you say "remember this:"
- **`memory`** - Simple memory management
- **`memory`** - Save and recall information (supports "remember this:", "save to memory", etc.)
**Community Skills (install via `/claude-os-skills`):**
- **36+ skills** from Anthropic Official and Superpowers repos
@ -600,7 +637,7 @@ Skills are reusable instruction sets that teach Claude specific capabilities. Th
**Global Skills** (`~/.claude/skills/`)
- Available in ALL projects
- Core skills: `memory`, `remember-this`, `initialize-project`
- Core skills: `memory`, `initialize-project`
**Project Skills** (`{project}/.claude/skills/`)
- Available only in that project
@ -842,7 +879,6 @@ claude-os/
│ │ └── ...
│ ├── skills/ # Global skills (symlinked to ~/.claude/)
│ │ ├── initialize-project/
│ │ ├── remember-this/
│ │ └── memory/
│ ├── skill-library/ # NEW: Local skill templates
│ │ ├── general/ # General purpose skills
@ -1093,7 +1129,6 @@ If you prefer to uninstall manually:
# Remove symlinks
rm ~/.claude/commands/claude-os-*.md
rm -rf ~/.claude/skills/initialize-project
rm -rf ~/.claude/skills/remember-this
rm -rf ~/.claude/skills/memory
rm ~/.claude/mcp-servers/code-forge.json

View file

@ -75,7 +75,7 @@ class SkillManager:
TEMPLATES_DIR = Path(__file__).parent.parent.parent / "templates" / "skill-library"
# Core skills that should remain global
CORE_SKILLS = {"memory", "remember-this"}
CORE_SKILLS = {"memory"}
def __init__(self, project_path: Optional[str] = None):
"""

View file

@ -48,7 +48,7 @@ CLAUDE_OS_COMMANDS=(
# List of Claude OS skills to move
CLAUDE_OS_SKILLS=(
"initialize-project"
"remember-this"
"memory"
"memory"
)
@ -131,8 +131,8 @@ if [ -f "$INIT_SKILL_MD" ]; then
echo " ✅ Updated initialize-project/SKILL.md"
fi
# Update paths in remember-this skill
REMEMBER_SKILL_MD="${TEMPLATES_DIR}/skills/remember-this/SKILL.md"
# Update paths in memory skill
REMEMBER_SKILL_MD="${TEMPLATES_DIR}/skills/memory/SKILL.md"
if [ -f "$REMEMBER_SKILL_MD" ]; then
# Backup original
cp "$REMEMBER_SKILL_MD" "${REMEMBER_SKILL_MD}.backup"
@ -140,7 +140,7 @@ if [ -f "$REMEMBER_SKILL_MD" ]; then
# Replace hardcoded user paths
sed -i.bak 's|'"${USER_CLAUDE_DIR}"'/|~/.claude/|g' "$REMEMBER_SKILL_MD"
rm -f "${REMEMBER_SKILL_MD}.bak"
echo " ✅ Updated remember-this/SKILL.md"
echo " ✅ Updated memory/SKILL.md"
fi
echo ""

View file

@ -154,7 +154,7 @@ Replace `20251101_135211` with your actual backup timestamp.
rm ~/.claude/commands/claude-os-*.md
rm ~/.claude/skills/memory
rm ~/.claude/skills/initialize-project
rm ~/.claude/skills/remember-this
rm ~/.claude/skills/memory
# Remove virtual environment
rm -rf venv venv_py312 venv_py313
@ -200,7 +200,7 @@ Replace `20251101_135211` with your actual backup timestamp.
# 2. Clean installation (keep data, test scripts)
rm -rf venv
rm ~/.claude/commands/claude-os-*.md
rm ~/.claude/skills/{memory,initialize-project,remember-this}
rm ~/.claude/skills/{memory,initialize-project,memory}
# 3. Test install as if you're a new user
./install.sh
@ -222,7 +222,7 @@ rm -rf venv venv_*
rm -rf node_modules
rm -rf data/claude-os.db
rm ~/.claude/commands/claude-os-*.md
rm ~/.claude/skills/{memory,initialize-project,remember-this}
rm ~/.claude/skills/{memory,initialize-project,memory}
# 3. Test full setup
./setup.sh

View file

@ -169,7 +169,7 @@ All these work in any initialized project:
## Skills Available
- `initialize-project` - Analyze codebase and generate standards
- `remember-this` - Auto-save when you say "remember this:"
- `memory` - Auto-save when you say "remember this:"
- `memory` - Simple memory management
## Agent-OS Commands (If Installed)

View file

@ -25,7 +25,7 @@ Available in ALL projects. These are core skills that come with Claude OS:
| Skill | Description |
|-------|-------------|
| `memory` | Save and recall information across sessions |
| `remember-this` | Auto-save on trigger phrases like "remember this:" |
| `memory` | Auto-save on trigger phrases like "remember this:" |
| `initialize-project` | Analyze codebase and generate documentation |
### Project Skills (`{project}/.claude/skills/`)
@ -71,7 +71,7 @@ The `/claude-os-skills` command provides full skills management:
🌐 GLOBAL SKILLS (always available)
────────────────────────────────────
✓ memory - Save and recall information
✓ remember-this - Auto-save on trigger phrases
✓ memory - Auto-save on trigger phrases
✓ initialize-project - Analyze codebase
📁 PROJECT SKILLS (/path/to/project)

View file

@ -79,7 +79,7 @@ fi
echo "" >> "$BACKUP_DIR/symlink_info.txt"
echo "# Symlinked Skills" >> "$BACKUP_DIR/symlink_info.txt"
if [ -d "$HOME/.claude/skills" ]; then
ls -la "$HOME/.claude/skills" | grep -E "(memory|initialize-project|remember-this)" >> "$BACKUP_DIR/symlink_info.txt" 2>/dev/null || echo "No claude-os skills found" >> "$BACKUP_DIR/symlink_info.txt"
ls -la "$HOME/.claude/skills" | grep -E "(memory|initialize-project|memory)" >> "$BACKUP_DIR/symlink_info.txt" 2>/dev/null || echo "No claude-os skills found" >> "$BACKUP_DIR/symlink_info.txt"
fi
echo -e "${GREEN}✓${NC} Recorded symlink information"

View file

@ -46,7 +46,7 @@ $([ -d "${CLAUDE_OS_DIR}/venv" ] && ${CLAUDE_OS_DIR}/venv/bin/pip list 2>&1 | gr
FILES CREATED
───────────────────────────────────────────────────────────────
Commands: $(ls -1 ${USER_CLAUDE_DIR}/commands/claude-os-*.md 2>/dev/null | wc -l | xargs) symlinks
Skills: $(ls -1d ${USER_CLAUDE_DIR}/skills/{memory,initialize-project,remember-this} 2>/dev/null | wc -l | xargs) symlinks
Skills: $(ls -1d ${USER_CLAUDE_DIR}/skills/{memory,initialize-project,memory} 2>/dev/null | wc -l | xargs) symlinks
start.sh: $([ -f "${CLAUDE_OS_DIR}/start.sh" ] && echo "Created" || echo "Not created")
config.json: $([ -f "${CLAUDE_OS_DIR}/claude-os-config.json" ] && echo "Created" || echo "Not created")
MCP config: $([ -f "${USER_CLAUDE_DIR}/mcp-servers/code-forge.json" ] && echo "Created" || echo "Not created")
@ -55,7 +55,7 @@ SYMLINK DETAILS
───────────────────────────────────────────────────────────────
$(ls -la ${USER_CLAUDE_DIR}/commands/claude-os-*.md 2>&1 | head -10)
$(ls -la ${USER_CLAUDE_DIR}/skills/{memory,initialize-project,remember-this} 2>&1 | head -5)
$(ls -la ${USER_CLAUDE_DIR}/skills/{memory,initialize-project,memory} 2>&1 | head -5)
GIT INFORMATION
───────────────────────────────────────────────────────────────

View file

@ -34,7 +34,7 @@ mcp__code-forge__list_skills
🌐 GLOBAL SKILLS (always available)
────────────────────────────────────
✓ memory - Save and recall information
✓ remember-this - Auto-save on trigger phrases
✓ memory - Auto-save on trigger phrases
📁 PROJECT SKILLS ({cwd})
────────────────────────────────────
@ -211,7 +211,7 @@ mcp__code-forge__delete_skill
✓ Deleted skill '<name>' from project
```
**Note:** Cannot delete global skills (memory, remember-this). These are core to Claude OS.
**Note:** Cannot delete global skills (memory, memory). These are core to Claude OS.
---
@ -219,7 +219,7 @@ mcp__code-forge__delete_skill
- **Global Skills**: `~/.claude/skills/`
- Available in ALL projects
- Core skills: memory, remember-this
- Core skills: memory, memory
- **Project Skills**: `{project}/.claude/skills/`
- Available only in that project

View file

@ -41,7 +41,7 @@ Claude OS Trigger Categories
============================
remember_this (ENABLED)
Invokes: remember-this skill
Invokes: memory skill
Phrases: 7 configured
Description: Auto-saves to Claude OS memories
@ -75,7 +75,7 @@ remember_this Triggers
======================
Status: ENABLED ✓
Invokes: remember-this skill
Invokes: memory skill
Case Sensitive: No
Trigger Phrases (7):
@ -88,12 +88,12 @@ Trigger Phrases (7):
7. "note:"
How it works:
When you say any of these phrases, the remember-this skill is automatically
When you say any of these phrases, the memory skill is automatically
invoked to save your content to Claude OS memories.
Examples:
User: "Remember this: We use service objects for business logic"
→ Auto-invokes remember-this skill
→ Auto-invokes memory skill
→ Saves to {project}-project_memories
To add/remove triggers:
@ -173,7 +173,7 @@ Input: "remember this: test phrase"
Matched Category: remember_this
Matched Phrase: "remember this:"
Would Invoke: remember-this skill
Would Invoke: memory skill
Status: ENABLED
Content to process: "test phrase"
@ -249,7 +249,7 @@ Use **Write** tool with complete JSON to update the file.
## Default Trigger Categories
### remember_this
Auto-invokes remember-this skill to save to memories
Auto-invokes memory skill to save to memories
- Default: ENABLED
- 7 default phrases
@ -302,7 +302,7 @@ Structure:
## Integration with Skills
The remember-this skill will be updated to:
The memory skill will be updated to:
1. Load triggers from this config file
2. Detect any matching phrase
3. Extract content after the trigger phrase

View file

@ -283,8 +283,7 @@ Parameters: kb_name (e.g., "{{PROJECT_NAME}}-project_memories"), query (your sea
### Available Skills
1. **`initialize-project`** - Analyze project and generate coding standards
2. **`remember-this`** - Save insights to Claude OS knowledge bases
3. **`memory`** - Simple memory management
2. **`memory`** - Save and recall information (supports "remember this:", "save to memory", etc.)
{{AGENT_OS_SECTION}}

View file

@ -1,134 +1,121 @@
---
name: memory
description: "Save and recall important information from previous sessions. Use when you want to remember technical insights, decisions, project context, or discoveries. Simply say 'remember:' or 'save to memory:' and I'll store it persistently and automatically reference it in future conversations."
allowed-tools: Read, Write, Bash
description: "Save and recall information across sessions. Use when you hear 'remember this', 'save to memory', 'add to your knowledge', or similar. Stores to Claude OS knowledge bases for persistent recall."
---
# Memory Skill
## Purpose
Create a persistent memory across all Claude Code sessions without needing to search through knowledge bases. Perfect for:
I use this skill to save important information to my Claude OS knowledge bases so I can recall it in future sessions. This is MY memory system - it makes me smarter over time.
- **Yesterday's work**: "What did we work on yesterday?"
- **Technical insights**: "Remember: Block rendering works like X"
- **Project context**: "Store: We're migrating Block 122 to Rails"
- **Debugging solutions**: "Save: Fixed N+1 query by adding .includes(:blocks)"
- **Architectural decisions**: "Remember: Always use Service classes for business logic"
- **Quick facts**: "Block 122 is 43KB, 1,150 lines"
## Trigger Phrases
## How to Use
When you say anything like:
- "remember this: ..."
- "save this to your memory"
- "add this to your knowledge"
- "don't forget that..."
- "store this: ..."
- "remember that..."
- "save to memory: ..."
- "keep this in mind: ..."
Simply say one of:
- **"Remember: [information]"**
- **"Save to memory: [information]"**
- **"Memory: [information]"**
- **"Store this: [information]"**
## What I Do
Or ask me to recall:
- **"What did we work on yesterday?"**
- **"Remind me about the Block deprecation"**
- **"Search my memory for 'Block 122'"**
- **"Show me my recent memories"**
1. **Extract** the key information from what you said
2. **Generate** a clear title and appropriate category
3. **Save** to `{project}-project_memories` knowledge base
4. **Confirm** briefly: "Saved: [title]"
## What Happens
No questions. No ceremony. Just save it.
When you invoke this skill:
## How I Save
1. **I extract key information** from your statement
2. **I store it in the Memory MCP** with tags and timestamp
3. **I confirm what was saved** and provide instant recall
4. **In future sessions**, I automatically reference this memory when relevant
I use the Claude OS API directly:
## Memory Structure
Each memory includes:
- **Title**: Quick reference name
- **Content**: Full details
- **Tags**: Keywords for searching (e.g., "Block122", "migration", "appointment-form")
- **Created**: Timestamp
- **Updated**: Last modification time
## Examples
### Example 1: Save Technical Insight
```
You: "Remember: The UserProfile component is the most complex in our app.
It has a 4-step wizard, 600+ lines of inline JavaScript, and handles
phone verification, profile lookup, and settings management."
→ I save to memory with tags: ["Block122", "Architecture", "Complexity"]
→ I confirm: "✓ Saved to memory: Block 122 complexity analysis"
```bash
curl -s -X POST "http://localhost:8051/api/kb/{project}-project_memories/upload" \
-F "title=[Generated Title]" \
-F "category=[Category]" \
-F "file=@/tmp/memory.md"
```
### Example 2: Project Progress
```
You: "Save to memory: We completed Block deprecation research.
Created BLOCKS_DEPRECATION_STRATEGY.md and case study for Block 122.
Next: Build feature flag infrastructure."
## Document Format
→ I save with tags: ["BlockDeprecation", "Progress", "Roadmap"]
→ In future session: "Based on our memory, we completed..."
```
```markdown
# [Title]
### Example 3: Recall with Context
```
You: "What was special about Block 122?"
→ I search memory automatically
→ "From our memory: Block 122 is a 4-step wizard form with 43KB content..."
```
## Key Benefits
✅ **Instant Recall** - No searching through KB or scrolling history
✅ **Automatic Context** - I know what we worked on without asking
✅ **Persistent** - Memories survive session restarts
✅ **Searchable** - Find memories by title, content, or tags
✅ **Organized** - See recent work, all memories, or search specific topics
✅ **Natural** - Just say "remember" and I handle the rest
## Memory Operations
### Add Memory
```
remember: [Your information here]
```
### Search Memory
```
search memory for: [keyword or phrase]
```
### Show Recent Memories
```
show my recent memories (from last 24 hours)
show my memories from yesterday
```
### Show All Memories
```
list all my memories
```
### Update Memory
```
update memory: [title or keyword] with [new information]
```
### Clear Memory (if needed)
```
forget: [memory title or keyword]
```
## Technical Details
**Storage**: ~/.claude/mcp-servers/memory/data/memories.json
**Searchable**: Full-text search on titles, content, and tags
**Timestamped**: All memories include creation and update times
**Persistent**: Survives across all sessions and devices
**Date Saved**: [YYYY-MM-DD]
**Category**: [Category]
---
**Pro Tip**: Use this liberally! Every insight, decision, and discovery you share builds your persistent context. I remember everything you tell me to remember, so you don't have to repeat yourself.
[The information you asked me to remember, well-structured]
## Key Points
- [Extracted key point 1]
- [Extracted key point 2]
---
*Saved to Claude OS*
```
## Categories
I auto-detect the category based on content:
| Category | When to Use |
|----------|-------------|
| Architecture | System design, structure decisions |
| Pattern | Code patterns, conventions, best practices |
| Troubleshooting | Bug fixes, solutions, workarounds |
| Decision | Why we chose X over Y |
| Integration | External APIs, third-party services |
| Business Logic | Domain rules, workflows |
| Context | Project background, user preferences |
## Examples
### Example 1: Quick Save
```
You: "remember this: the auth system uses JWT with 15min access tokens and 7-day refresh tokens"
Me: Saved: "Authentication Token Strategy" (Architecture)
```
### Example 2: Pattern
```
You: "add to your knowledge - when creating services, always return the model on success or an error string on failure"
Me: Saved: "Service Return Pattern" (Pattern)
```
### Example 3: Troubleshooting
```
You: "don't forget that Rails 4 doesn't support the hash syntax for exists?"
Me: Saved: "Rails 4 ActiveRecord Compatibility" (Troubleshooting)
```
## Recall
When you ask me to recall, I search my knowledge base:
- "What do you remember about auth?"
- "Search your memory for service patterns"
- "What did we decide about the database?"
I use `mcp__code-forge__search_knowledge_base` to find relevant memories.
## Why This Matters
Every memory makes me better:
- I don't start cold next session
- I remember YOUR patterns and preferences
- I learn from past solutions
- I build institutional knowledge
**Use liberally. Every insight saved is an insight I'll have forever.**

View file

@ -1,107 +0,0 @@
---
name: remember-this
description: "Capture conversation context, insights, or discoveries and save them as markdown documents to Claude OS knowledge bases. Use when you want to save important information like 'remember this:', design decisions, troubleshooting solutions, or architectural patterns for future reference."
allowed-tools: Read, Write, Bash, Task
---
# Remember This Skill
## Purpose
Automatically capture important context from our conversations and save it to Claude OS knowledge bases so I can reference it in future sessions. Perfect for:
- Design decisions and rationale
- Troubleshooting solutions discovered
- Architectural patterns you've explained
- Edge cases encountered
- Integration patterns
- Best practices you've shared
- Complex system behavior
## How to Use
Simply say one of:
- **"Remember this: [information]"**
- **"Save this: [information]"**
- **"Document this: [information]"**
- **"Remember: [explanation]"**
## What Happens
When you invoke this skill:
1. **I extract the key information** from your statement
2. **I create a formatted markdown document** with proper structure
3. **I upload it to Claude OS** (default: {project}-project_memories KB)
4. **I confirm what was saved** and where
5. **I reference it in future conversations** automatically
## Document Structure
```markdown
# [Topic Title]
**Date Saved**: [Auto-generated]
**Context**: [Where this came from]
**Category**: [Architecture/Integration/Pattern/Troubleshooting/etc]
[Main content in well-structured format]
## Key Takeaways
- Bullet point 1
- Bullet point 2
## Related Topics
- Links to related documentation
```
## Examples
### Example 1: Remember a Pattern
```
User: "Remember this: When creating a new service, always follow this structure:
initialize with domain objects and params, have a single perform method that returns
model on success or error string on failure, use private methods for complex logic."
→ Saves to: "Service Pattern Guide.md" in project KB
→ I reference it: "Like the pattern you documented, services should..."
```
### Example 2: Save a Solution
```
User: "Save this: Fixed the N+1 query problem in AppointmentsScheduler by adding
.includes(:block_settings, :children) to prevent multiple DB queries."
→ Saves to: "Query Optimization Solutions.md"
→ I use it: "We can apply the same .includes strategy here..."
```
### Example 3: Document an Architecture Decision
```
User: "Remember: Block model uses acts_as_tree for hierarchical structure because
we need semantic role-based organization of child blocks, not just a flat list."
→ Saves to: "Blocks System Decisions.md"
→ I reference it: "Following the hierarchical pattern you documented..."
```
## Customization
You can specify:
- **KB Name**: "Remember this [content] - save to my-project-memories"
- **Document Title**: "Remember this [content] - as 'My Custom Title'"
- **Tags/Category**: "Remember this [content] - category: Integration"
## Benefits
✅ **Build institutional knowledge** over time
✅ **I remember context** across sessions
✅ **Reduces repetition** of explanations
✅ **Creates searchable archive** of decisions
✅ **Helps with onboarding** (your knowledge is documented)
✅ **Perfect for discoveries** made during debugging
---
**Pro Tip**: Use this skill liberally! Every insight you share gets permanently indexed and makes me a better developer for your project.

View file

@ -1,65 +0,0 @@
#!/bin/bash
# save-to-mcp.sh - Save context to Claude OS
# Usage: save-to-mcp.sh "Document Title" "Content" "KB Name (optional)" "Category (optional)"
TITLE="${1:-Captured Context}"
CONTENT="${2:-}"
KB_NAME="${3:-project_memories}"
CATEGORY="${4:-General}"
# Sanitize KB name for URL
KB_ENCODED=$(echo "$KB_NAME" | sed 's/ /%20/g')
# Create timestamp
TIMESTAMP=$(date -u +"%Y-%m-%dT%H:%M:%SZ")
DATE=$(date +"%Y-%m-%d")
# Create document
DOC_CONTENT=$(cat <<EOF
# $TITLE
**Date Saved**: $DATE
**Time**: $TIMESTAMP
**Category**: $CATEGORY
---
## Context
$CONTENT
---
*Saved to Claude OS - Your AI Memory System*
EOF
)
# Save locally first
DOC_FILENAME=$(echo "$TITLE" | sed 's/[^a-zA-Z0-9]/_/g').md
DOC_PATH="/tmp/${DOC_FILENAME}"
echo "$DOC_CONTENT" > "$DOC_PATH"
# Upload to Claude OS
API_URL="http://localhost:8051"
RESPONSE=$(curl -s -X POST \
"$API_URL/api/kb/$KB_ENCODED/upload" \
-F "file=@$DOC_PATH" \
-w "\n%{http_code}")
HTTP_CODE=$(echo "$RESPONSE" | tail -n 1)
# Check response
if [[ "$HTTP_CODE" =~ ^(200|201)$ ]]; then
echo "✅ Saved to Claude OS"
echo "📁 KB: $KB_NAME"
echo "📄 Title: $TITLE"
echo "🏷️ Category: $CATEGORY"
echo "💾 Local: $DOC_PATH"
exit 0
else
echo "❌ Failed to save (HTTP $HTTP_CODE)"
echo "Response: $(echo "$RESPONSE" | sed '$d')"
exit 1
fi

View file

@ -57,7 +57,7 @@ done
# Remove skill symlinks
echo ""
echo "📁 Removing skill symlinks..."
for skill in "initialize-project" "remember-this" "memory"; do
for skill in "initialize-project" "memory"; do
skill_path="${USER_CLAUDE_DIR}/skills/${skill}"
if [ -L "$skill_path" ]; then
rm -f "$skill_path"