Skill AWS Knowledge Tool

When to use

• When you need to search AWS documentation programmatically
• When you need to read and convert AWS docs to markdown
• When you need to discover related AWS documentation

AWS Knowledge Tool Skill

Purpose

This skill provides access to the aws-knowledge-tool CLI - a command-line interface for querying the AWS Knowledge MCP Server. Use it to search, read, and discover AWS documentation programmatically.

When to Use This Skill

Use this skill when:

• You need to search AWS documentation with full-text search
• You need to read AWS documentation pages as markdown
• You need to discover related documentation recommendations
• You need structured/JSON output for processing

Do NOT use this skill for:

• Writing new AWS documentation (read-only)
• Deploying AWS resources (documentation only)
• AWS CLI operations (use AWS CLI instead)

CLI Tool: aws-knowledge-tool

The aws-knowledge-tool is a CLI tool for querying the AWS Knowledge MCP Server to search, read, and discover AWS documentation.

Installation

# Clone and install
git clone https://github.com/dnvriend/aws-knowledge-tool.git
cd aws-knowledge-tool
uv tool install .

Prerequisites

• Python 3.14+
• uv package manager

Quick Start

# Search AWS documentation
aws-knowledge-tool search "Lambda function URLs"

# Read documentation page
aws-knowledge-tool read "https://docs.aws.amazon.com/lambda/latest/dg/welcome.html"

# Get recommendations
aws-knowledge-tool recommend "https://docs.aws.amazon.com/lambda/latest/dg/welcome.html"

Progressive Disclosure

search - Search AWS Documentation

Search across AWS documentation, blogs, solutions library, architecture center, and prescriptive guidance.

Usage:

aws-knowledge-tool search QUERY [OPTIONS]

Arguments:

• QUERY: Search query text (required)
• --limit N / -l N: Maximum results (default: 10)
• --offset M / -o M: Skip first M results for pagination (default: 0)
• --json: Output JSON format for processing
• --stdin: Read query from stdin (for pipelines)
• -v/-vv/-vvv: Verbosity (INFO/DEBUG/TRACE)

Examples:

# Basic search
aws-knowledge-tool search "S3 versioning"

# With pagination
aws-knowledge-tool search "Lambda" --limit 10 --offset 20

# JSON output
aws-knowledge-tool search "DynamoDB" --json

# Pipeline usage
echo "CloudFormation" | aws-knowledge-tool search --stdin --json

Output: Returns list of results with: title, url, context, rank_order

read - Read AWS Documentation

Fetch and convert AWS documentation pages to markdown format.

Usage:

aws-knowledge-tool read URL [OPTIONS]

Arguments:

• URL: AWS documentation URL (required)
• --start-index N / -s N: Starting character index for pagination
• --max-length M / -m M: Maximum characters to fetch
• --json: Output JSON format
• --stdin: Read URL from stdin
• -v/-vv/-vvv: Verbosity (INFO/DEBUG/TRACE)

Examples:

# Read full document
aws-knowledge-tool read "https://docs.aws.amazon.com/lambda/latest/dg/welcome.html"

# With pagination (large docs)
aws-knowledge-tool read "https://docs.aws.amazon.com/..." \
  --start-index 5000 --max-length 2000

# Pipeline from search
aws-knowledge-tool search "Lambda" --json | \
  jq -r '.[0].url' | \
  aws-knowledge-tool read --stdin

Output: Returns markdown-formatted documentation content.

Supported domains:

• docs.aws.amazon.com
• aws.amazon.com

recommend - Get Documentation Recommendations

Discover related documentation through four recommendation types.

Usage:

aws-knowledge-tool recommend URL [OPTIONS]

Arguments:

• URL: AWS documentation URL (required)
• --type TYPE / -t TYPE: Filter by recommendation type
  • highly_rated: Popular pages within same AWS service
  • new: Recently added pages (useful for finding new features)
  • similar: Pages covering similar topics
  • journey: Pages commonly viewed next by other users
• --limit N / -l N: Max results per category (default: 5)
• --offset M / -o M: Skip first M per category (default: 0)
• --json: Output JSON format
• --stdin: Read URL from stdin
• -v/-vv/-vvv: Verbosity (INFO/DEBUG/TRACE)

Examples:

# Get all recommendations
aws-knowledge-tool recommend "https://docs.aws.amazon.com/lambda/latest/dg/welcome.html"

# Filter by type (find new features)
aws-knowledge-tool recommend "https://docs.aws.amazon.com/..." --type new

# JSON output with limit
aws-knowledge-tool recommend "https://docs.aws.amazon.com/..." --json --limit 3

# Pipeline usage
aws-knowledge-tool search "Lambda" --json | \
  jq -r '.[0].url' | \
  aws-knowledge-tool recommend --stdin --type similar

Output: Returns dict with recommendation categories and their pages (title, url, context).

Multi-Level Verbosity Logging

Control logging detail with progressive verbosity levels. All logs output to stderr.

Logging Levels:

Examples:

# INFO level - see operations
aws-knowledge-tool search "Lambda" -v

# DEBUG level - see detailed info
aws-knowledge-tool search "Lambda" -vv

# TRACE level - see all internals
aws-knowledge-tool search "Lambda" -vvv

Shell Completion

Native shell completion for bash, zsh, and fish.

Installation:

# Bash (add to ~/.bashrc)
eval "$(aws-knowledge-tool completion bash)"

# Zsh (add to ~/.zshrc)
eval "$(aws-knowledge-tool completion zsh)"

# Fish (save to completions)
aws-knowledge-tool completion fish > ~/.config/fish/completions/aws-knowledge-tool.fish

Pipeline Composition

Compose commands with Unix pipes for powerful workflows.

Examples:

# Search → Extract URL → Read
aws-knowledge-tool search "Lambda" --json | \
  jq -r '.[0].url' | \
  aws-knowledge-tool read --stdin

# Search → Extract URL → Get similar docs
aws-knowledge-tool search "S3" --json | \
  jq -r '.[0].url' | \
  aws-knowledge-tool recommend --stdin --type similar

# Save search results to file
aws-knowledge-tool search "DynamoDB" --json > results.json

# Read and save as markdown
aws-knowledge-tool read "https://docs.aws.amazon.com/..." > lambda-docs.md

Common Issues

Issue: Command not found

# Verify installation
aws-knowledge-tool --version

# Reinstall if needed
cd aws-knowledge-tool
uv tool install . --reinstall

Issue: No results from search

• Check your search query is specific enough
• Try broader search terms
• Use --json to see full response

Issue: URL validation error

• Ensure URL is from docs.aws.amazon.com or aws.amazon.com
• Check URL is accessible in browser first

Issue: Connection timeout

• Check internet connection
• AWS Knowledge MCP Server may be temporarily unavailable
• Try again with verbose flag: -vv

Getting Help

# Show help
aws-knowledge-tool --help

# Command-specific help
aws-knowledge-tool search --help
aws-knowledge-tool read --help
aws-knowledge-tool recommend --help

Exit Codes

• 0: Success
• 1: Client error (invalid arguments, validation failed)
• 2: Server error (API error, network issue)
• 3: Network error (connection failed, timeout)

Output Formats

Default (Markdown):

• Human-readable formatted output
• Search: ranked list with titles, URLs, context
• Read: markdown content
• Recommend: grouped by category with titles, URLs

JSON (--json flag):

• Machine-readable structured data
• Perfect for pipelines and processing
• Consistent structure across commands

Best Practices

1. Use JSON for pipelines: Enable --json when piping to other tools
2. Pagination for large results: Use --limit and --offset for controlled fetching
3. Progressive verbosity: Start with -v, increase to -vv/-vvv only if needed
4. Save frequent searches: Cache JSON results to avoid repeated API calls
5. Compose commands: Leverage Unix pipes for powerful workflows

Resources

• GitHub: https://github.com/dnvriend/aws-knowledge-tool
• AWS Knowledge MCP Server: https://knowledge-mcp.global.api.aws
• MCP Specification: https://modelcontextprotocol.io