Security Considerations

- Define CLI commands as MCP tools using <<parameter>> placeholders in YAML
- Create reusable AI prompt templates with {{argument}} Handlebars syntax
- Use --config-path, --config-value, or MCP_THIS_CONFIG_PATH for configuration
- Pre-built presets: default (read-only), editing (file manipulation), github (GitHub integration)
- Optional parameters are automatically omitted from commands if not provided
- Supports conditional sections in prompts with {{#if}} blocks

Create my-tools.yaml:

tools:
  web-scraper:
    description: Fetch a webpage and convert it to clean, readable text
    execution:
      command: curl -s '<<url>>' | lynx -dump -stdin
    parameters:
      url:
        description: URL of the webpage to fetch
        required: true
find-large-files:

    description: Find files larger than specified size in a directory

    execution:

      command: find '<<directory>>' -type f -size +<<size>> -exec ls -lh {} \;

    parameters:

      directory:

        description: Directory to search

        required: true

      size:

        description: Minimum file size (e.g., 100M, 1G)

        required: true

prompts:
summarize-webpage:
description: Generate a structured summary of webpage content
template: |
Please analyze the following webpage content and provide:

1. Main Topic: What is this page about?
2. Key Points: {{num_points}} most important points
3. Target Audience: Who is this content for?
{{#if focus}}4. {{focus}} Analysis: Specific insights about {{focus}}{{/if}}

Content:
{{content}}
arguments:
content:
description: Webpage content to summarize
required: true
num_points:
description: Number of key points to extract (default 5)
required: false
focus:
description: Specific aspect to focus on (e.g., technical, business, educational)
required: false

file-analysis:
description: Analyze files for specific purposes
template: |
Analyze the following files for {{analysis_type}}:

{{#if criteria}}Focus on: {{criteria}}{{/if}}

{{files}}

Please provide:
- Summary of findings
- Recommendations
- {{#if format}}Output in {{format}} format{{/if}}
arguments:
files:
description: File contents or paths to analyze
required: true
analysis_type:
description: Type of analysis (security, performance, quality, etc.)
required: true
criteria:
description: Specific criteria or standards to check against
required: false
format:
description: Output format (markdown, JSON, report, etc.)
required: false

Using Prompts in Claude Desktop:

1. Click the + icon in the message input

2. Select "Add from mcp-this-custom"

3. Choose your prompt (e.g., "summarize-webpage")

4. Fill in the arguments - Claude will guide you through the required and optional fields

Key Points:

- Use <<parameter>> placeholders in commands

- Parameters marked required: false are removed from commands if not provided

- Use command: >- for multi-line commands (not command: |)

For convenience, mcp-this includes ready-to-use collections of tools and prompts:
- default - Safe, read-only tools (file exploration, web scraping)

- editing - File manipulation tools (create, edit, delete)  

- github - GitHub integration tools (PR analysis, repository operations) + specialized prompts (code-review, create-pr-description)
Quick usage:

> See README_PRESETS.md for complete preset documentation, tool lists, dependencies, and advanced setup.

---

test-runner:
description: Run tests with optional pattern matching
execution:
command: >-
if [ -n "<<pattern>>" ]; then
npm test -- --grep "<<pattern>>"
else
npm test
fi
parameters:
pattern:
description: Test pattern to match (optional)
required: false

docker-container-logs:
description: Get logs from a Docker container
execution:
command: docker logs <<container_name>> --tail <<lines>>
parameters:
container_name:
description: Name or ID of the Docker container
required: true
lines:
description: Number of log lines to show (default 100)
required: false
default: "100"

service-status:
description: Check the status of a system service
execution:
command: systemctl status <<service_name>>
parameters:
service_name:
description: Name of the service to check
required: true

disk-usage-analyzer:
description: Analyze disk usage and find largest directories
execution:
command: >-
echo "=== Disk Usage Summary ===" &&
df -h <<path>> &&
echo "=== Largest Directories ===" &&
du -h <<path>> | sort -hr | head -10
parameters:
path:
description: Path to analyze (default current directory)
required: false
default: "."

---

tools = await session.list_tools()

        print([tool.name for tool in tools.tools])
result = await session.call_tool(

            'git-status-summary',

            {}

        )

        print(result.content[0].text)

---

uvx mcp-this --config-path ./my-tools.yaml
```