count-items
Count items in a DynamoDB table
LABSCO SUMMARY
The tool list contains no write operation, so what the agent can do is bounded by the tools even where the AWS credentials you hand it are broader.
The distinction that decides your bill is query-table against scan-table: a query is bounded by keyConditionExpression and can be pointed at a secondary index through indexName, while a scan reads the table and applies filterExpression afterwards. count-items answers a how-many question without moving the items across the wire. paginate-query-table takes a narrower argument list than query-table — no indexName, no limit, no filterExpression — so a paged read gives up the index target and the filter that the single-shot query accepts.
What it is
A read-only DynamoDB client exposing list-tables, describe-table, get-item, query-table, paginate-query-table, scan-table and count-items.
What you get
- Discovery before access:
list-tablesreturns the table names in the account, anddescribe-tablereturns the detail of a single table bytableName. - Key-addressed reads through
get-item, which takes thekeyand thetableNameand returns that item. - Condition-bounded reads through
query-table, which takes akeyConditionExpressionwithexpressionAttributeValues, can be aimed at a secondary index viaindexName, and narrows the result withfilterExpression,projectionExpressionandlimit. - Whole-table reads through
scan-table, under the samefilterExpression,projectionExpressionandlimitcontrols. paginate-query-tablefor a query whose result set runs past a single response.count-items, which returns how many items match afilterExpressionwithout returning the items themselves.
Requirements
AWS credentials for the account holding the tables — AWS_ACCESS_KEY_ID and AWS_SECRET_ACCESS_KEY — plus the region those tables live in as AWS_REGION.
Setup effort
One command plus a key — npx -y @smithery/cli install @jjikky/dynamo-readonly-mcp --client claude, then supply credentials
USE CASES
01
Answering a question about production data without holding write access
Find the table, read its schema to learn the key structure, then run a condition-bounded query against it.
02
Sizing a result set before you commit to paging through it
Count what the filter matches first, then page the query only if the number justifies it.
03
Pulling a single record when you already hold its key
Skip the query planning entirely and address the item directly.
04
Hunting for rows on an attribute the table is not keyed by
Read the schema to confirm no index covers the attribute, then filter over a scan and keep the projection tight.
7 tools
Counted once with only startup setup. Other versions, modes or optional credentials may expose a different list.
describe-table
Get detailed information about a DynamoDB table
get-item
Get an item from a DynamoDB table based on a specific key
list-tables
Get a list of all DynamoDB tables
paginate-query-table
Paginate query results
query-table
Query items from a DynamoDB table based on conditions
scan-table
Scan items from a DynamoDB table
BEFORE YOU INSTALL
What to know first
Tools you get
7 tools when we connected. Every tool reads, and no flag enables writes.
Requires
- npx on PATH
- AWS_ACCESS_KEY_ID in your environment
- AWS_SECRET_ACCESS_KEY in your environment
CONTEXT COST
Light Standing cost
Rated by the size of the 7 tool definitions that ride along on every turn, not by how many there are
INSTALLATION
Connect your client
Server setup
Smithery
TERMINAL
npx -y @smithery/cli install @jjikky/dynamo-readonly-mcp --client claudeMCP Clients
Claude Desktop
CLAUDE DESKTOP · JSON
{
"mcpServers": {
"dynamodb-readonly": {
"command": "npx",
"args": [
"-y",
"dynamo-readonly-mcp"
],
"env": {
"AWS_ACCESS_KEY_ID": "<YOUR_AWS_ACCESS_KEY_ID>",
"AWS_SECRET_ACCESS_KEY": "<YOUR_AWS_SECRET_ACCESS_KEY>",
"AWS_REGION": "<YOUR_AWS_REGION>"
}
}
}
}
