🔥 chore(skills): remove deprecated mcp-memorygraph and mcp-ncp skill documentation

---

name: mcp-memorygraph

description: Store and recall project-specific context, decisions, and patterns across coding sessions using Memory-Graph

---

# Memory Protocol (Memory-Graph)

**If MCP server `memorygraph` is available**, use it to store and retrieve project context automatically.

## REQUIRED: Before Starting Work

You MUST use `recall_memories` before any task to avoid repeating work.

```python

recall_memories(query="tdoc-crawler crawling and database patterns")

```

## REQUIRED: Automatic Storage Triggers

Store memories on ANY of:

- **Git commit** → What was fixed/added

- **Bug fix** → Problem + solution

- **Version release** → Summarize changes

- **Architecture decision** → Choice + rationale

- **Pattern discovered** → Reusable approach

## Timing Mode

```python

memory_mode: "immediate" | "on-commit" | "session-end"

```

- **immediate**: Store immediately after action

- **on-commit**: Trigger on git commits (default)

- **session-end**: Store at end of coding session

## Memory Fields

```python

{

  "type": "solution" | "problem" | "code_pattern" | "fix" | "error" | "workflow",

  "title": "Specific, searchable (not generic)",

  "content": "Accomplishment, decisions, patterns",

  "tags": ["project", "tech", "category"],  # REQUIRED

  "importance": 0.8,  # Critical (0.8+), Standard (0.5-0.7), Minor (0.3-0.4)

  "relationships": []  # Link related memories when they exist

}

```

## Tag Guidelines

Tags are REQUIRED for effective retrieval:

- **project**: Project name or identifier

- **tech**: Technology stack component (e.g., "pydantic", "sqlite", "pytest")

- **category**: Type of memory (e.g., "database", "authentication", "testing")

## Best Practices

- **Don't wait to be asked**: Memory storage is automatic

- **Be specific**: Titles should be searchable, not generic

- **Link related memories**: Connect related decisions/patterns

- **Use proper type**: Choose correct type for each memory

- **Set appropriate importance**: Help prioritize retrieval

## Query Examples

```python

# Recall by project

recall_memories(query="tdoc-crawler database")

# Recall by technology

recall_memories(query="pydantic models")

# Recall by pattern

recall_memories(query="caching patterns")

# Recall by specific task

recall_memories(query="TDoc authentication")

```

## References

- [Memory-Graph GitHub](https://github.com/gregorydickson/memory-graph)

---

name: mcp-ncp

description: Tool discovery and delegation to appropriate MCP servers based on task analysis

---

# NCP (Natural Context Protocol)

**If MCP server `ncp` is available**, use it to discover the most suitable tools and MCP servers for your current task.

## When to Use

Use ncp BEFORE starting significant work to:

- Analyze task description

- Suggest appropriate tools

- Delegate to MCP servers

- Find optimal solutions

## Workflow

1. **Provide task**: Describe what you need to do

2. **Get recommendations**: ncp analyzes and suggests tools

3. **Follow delegation**: ncp may delegate to other MCP servers

4. **Execute solution**: Use the recommended tool/server

## Example

When you need to analyze a problem:

```python

# Describe your task

"I need to analyze test failures in the pytest suite"

# NCP may suggest:

# - pytest analysis tools

# - debugging MCP servers

# - test visualization tools

```

## Best Practices

- **Use before starting**: Discover tools before beginning work

- **Follow recommendations**: Trust ncp's analysis for tool selection

- **Consider context**: ncp considers available MCP servers

## References

- [NCP GitHub](https://github.com/porfell-dev/ncp)