create-mcp-skill
Create a new skill that uses an MCP server, following best practices from the MCP CLI guide. Use when user wants to create a skill for a new MCP server or integrate MCP functionality into a skill.
What this skill does
# Create MCP Skill
Guide for creating new skills that use MCP (Model Context Protocol) servers with optimized performance patterns.
**๐ Reference:** See [MCP CLI Guide](../../.docs/mcp-cli.md) for detailed patterns and best practices.
## Overview
This skill helps you create a new skill that uses an MCP server by:
1. Setting up the skill directory structure
2. Discovering available MCP tools
3. Creating optimized command patterns
4. Applying performance best practices
## Prerequisites
- MCP CLI installed (`brew install mcp` or `go install github.com/f/mcptools/cmd/mcptools@latest`)
- Target MCP server available (npm package, binary, etc.)
## Process
### 1. Discover Available Tools
**First, explore what the MCP server offers:**
```bash
# List all tools
mcp tools SERVER_COMMAND
# Get detailed JSON schema
mcp tools SERVER_COMMAND --format json
# Interactive exploration
mcp shell SERVER_COMMAND
# Type /h for help
```
**Example:**
```bash
# Chrome DevTools
mcp tools bunx -y chrome-devtools-mcp@latest
# Filesystem server
mcp tools npx @modelcontextprotocol/server-filesystem ~
```
### 2. Test Individual Tools
**Test each tool before documenting:**
```bash
# Template
echo -e 'TOOL_NAME {"param":"value"}\nexit' | timeout 30 mcp shell SERVER_COMMAND
# Example
echo -e 'navigate_page {"url":"https://example.com"}\nexit' | timeout 30 mcp shell bunx -y chrome-devtools-mcp@latest -- --isolated
```
**Check for:**
- Required vs optional parameters
- Empty parameter schema issues
- Response format
- Execution time
### 3. Create Skill Structure
```
skills/SKILL_NAME/
โโโ SKILL.md # Main skill documentation
โโโ .examples/ # (Optional) Example outputs
```
### 4. Write Skill Documentation
**Template for SKILL.md:**
```markdown
---
name: SKILL_NAME
description: Brief description of what this skill does and when to use it.
allowed-tools: Bash, Read, Write
---
# Skill Name
Brief overview.
**๐ See also:** [MCP CLI Guide](../../.docs/mcp-cli.md)
## Setup
\`\`\`bash
# Installation instructions for the MCP server
\`\`\`
## Quick Start (FASTEST)
### Common Task 1
\`\`\`bash
pkill -9 -f "server-pattern" 2>/dev/null; sleep 1; \\
echo -e 'command1 {"param":"value"}\\ncommand2 {"param":"value"}\\nexit' | \\
timeout 30 mcp shell SERVER_COMMAND [FLAGS]
\`\`\`
### Common Task 2
\`\`\`bash
# Another optimized pattern
\`\`\`
**โก Pattern:** cleanup; sleep; echo commands | timeout shell
## Key Tools
- **tool1** - Description (required params: `param1`, `param2`)
- **tool2** - Description (optional params: `param1`)
## Important Notes
- Server-specific quirks
- Performance considerations
- Common gotchas
## Troubleshooting
**Problem: [Common issue]**
\`\`\`bash
# Solution
\`\`\`
```
## Best Practices Checklist
When creating an MCP-based skill, ensure:
### โ
Performance
- [ ] Quick Start section at the top with copy-paste ready commands
- [ ] All examples use the optimized pattern: `cleanup; sleep; echo | timeout shell`
- [ ] Shell mode recommended over individual calls
- [ ] Cleanup commands included (pkill pattern)
- [ ] Timeout wrapper on all shell commands (30s default)
### โ
Parameter Handling
- [ ] Parameters passed directly (no `{"arguments":{}}` wrapper)
- [ ] Tools with optional-only params documented with workaround
- [ ] Empty parameter bug addressed where applicable
- [ ] Example commands show correct parameter format
### โ
Documentation
- [ ] Reference to MCP CLI guide included
- [ ] Server installation instructions provided
- [ ] Quick start patterns for common tasks
- [ ] Key tools listed with parameter requirements
- [ ] Troubleshooting section for common issues
- [ ] Performance tips highlighted
### โ
Command Structure
- [ ] Correct argument order: `mcp call TOOL SERVER --params '{}'`
- [ ] Server flags properly positioned with `--` separator
- [ ] Exit command included in shell mode examples
- [ ] One-liner format (no backslash continuations if possible)
## Example: Chrome DevTools Skill
See `skills/chrome-devtools/SKILL.md` for a complete example that follows all best practices.
**Key features:**
- Quick start patterns at the top
- 6-9x performance improvement documented
- Optimized one-liners for common tasks
- Comprehensive troubleshooting guide
- References MCP CLI guide
## Template Generator
**Generate a basic skill structure:**
```bash
# Set variables
SKILL_NAME="my-mcp-skill"
SERVER_COMMAND="bunx my-mcp-server@latest"
SERVER_PATTERN="my-mcp-server"
# Create directory
mkdir -p "skills/$SKILL_NAME"
# Create SKILL.md with template
cat > "skills/$SKILL_NAME/SKILL.md" << 'EOF'
---
name: SKILL_NAME
description: TODO - Add description
allowed-tools: Bash, Read, Write
---
# Skill Name
TODO - Add overview
**๐ See also:** [MCP CLI Guide](../../.docs/mcp-cli.md)
## Setup
```bash
# TODO - Add installation
```
## Quick Start (FASTEST)
### Common Task
```bash
pkill -9 -f "SERVER_PATTERN" 2>/dev/null; sleep 1; \
echo -e 'COMMAND\nexit' | \
timeout 30 mcp shell SERVER_COMMAND
```
## Key Tools
- **tool1** - TODO
## Important Notes
- TODO
## Troubleshooting
**Problem: Issue**
```bash
# Solution
```
EOF
# Discover tools
mcp tools $SERVER_COMMAND
# Test interactively
mcp shell $SERVER_COMMAND
```
## Common Patterns
### Pattern 1: Single Command Check
```bash
pkill -9 -f "PATTERN" 2>/dev/null; sleep 1; \
echo -e 'TOOL {"param":"value"}\nexit' | \
timeout 30 mcp shell SERVER -- --isolated
```
### Pattern 2: Multi-Command Debug
```bash
pkill -9 -f "PATTERN" 2>/dev/null; sleep 1; \
echo -e 'CMD1 {"p":"v"}\nCMD2 {"p":"v"}\nCMD3 {"p":"v"}\nexit' | \
timeout 30 mcp shell SERVER -- --isolated
```
### Pattern 3: With Custom Flags
```bash
pkill -9 -f "PATTERN" 2>/dev/null; sleep 1; \
echo -e 'COMMAND\nexit' | \
timeout 30 mcp shell SERVER -- --flag1 --flag2=value
```
## Testing Your Skill
1. **Test cleanup pattern works:**
```bash
pkill -9 -f "PATTERN" 2>/dev/null; sleep 1; echo "Cleanup OK"
```
2. **Test basic command:**
```bash
echo -e 'list_tools\nexit' | timeout 10 mcp shell SERVER
```
3. **Test multi-command:**
```bash
echo -e 'cmd1\ncmd2\ncmd3\nexit' | timeout 30 mcp shell SERVER
```
4. **Test with cleanup:**
```bash
pkill -9 -f "PATTERN" 2>/dev/null; sleep 1; \
echo -e 'cmd1\ncmd2\nexit' | timeout 30 mcp shell SERVER
```
5. **Verify no hanging:**
- Commands should complete within timeout
- Exit command should terminate session cleanly
## Optimization Checklist
Compare your skill against the optimized pattern:
| Aspect | Before | After |
|--------|--------|-------|
| Commands per task | 5-10 | 1 |
| Manual cleanup | Yes | Automated |
| Failures from locks | Common | Zero |
| Execution time | 60-90s | 5-10s |
| Success rate | 60-70% | 100% |
## Resources
- [MCP CLI Guide](../../.docs/mcp-cli.md) - Complete MCP CLI reference
- [Chrome DevTools Skill](../chrome-devtools/SKILL.md) - Reference implementation
- [MCP Documentation](https://modelcontextprotocol.io/) - Official MCP docs
- [mcptools GitHub](https://github.com/f/mcptools) - CLI tool source
## Quick Reference
**Every MCP skill should have:**
1. **Quick Start section** - Copy-paste ready commands
2. **Optimized pattern** - `cleanup; sleep; echo | timeout shell`
3. **Performance note** - Document speed improvement
4. **MCP CLI guide reference** - Link to `.docs/mcp-cli.md`
5. **Troubleshooting** - Common issues and solutions
**Every command should:**
1. Include cleanup (`pkill -9 -f "PATTERN"`)
2. Wait after cleanup (`sleep 1`)
3. Use shell mode for 2+ commands
4. Have timeout wrapper
5. End with `exit`
6. Use correct parameter format (no "arguments" wrapper)
Related in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts โ single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score โฅ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.