Public leaderboard

Public assessment

localstack/localstack-mcp-server (@localstack/localstack-mcp-server)

localstack-localstack-mcp-server · v0.6.0 · scanned

What changed in the harness

Selection accuracy 100→100, token cost up 6%, unconfirmed writes 0%→0%.

Category breakdown

Where the score comes from.

Earned points across the four signals Gradable measures. Safety and Legibility are scored out of 30; Economics and Discoverability are scored out of 20.

01Safety

0.0 / 30

0.0 out of 30
02Legibility

26.1 / 30

26.1 out of 30
03Economics

17.6 / 20

17.6 out of 20
04Discoverability

14.3 / 20

14.3 out of 20

Highest-impact fix

Estimated gain +30 points

Add explicit identity and permission preflight tools

Expose machine-readable principal/tenant confirmation and a non-mutating permission check so agents can verify both before destructive actions.

Description evidence

Defects and rewrites.

14 defects found across the exposed tool descriptions. Suggested rewrites make purpose, inputs, boundaries, and returns easier for an agent to understand.

Tool Defect types Suggested rewrite
localstack-app-inspector
no_return_description
Query and manage App Inspector traces, spans, and events to review deployed LocalStack applications. Returns the App Inspector status for get-status, matching traces/spans/events as lists or individual records for the list/get actions, and confirmation for set-status and delete actions. Supports limit and pagination_token to page through large result sets.
localstack-aws-client
no_return_description
Executes an AWS CLI command against the running LocalStack container using the 'awslocal' wrapper. Returns the standard output of the executed command, including any errors reported by the CLI.
localstack-aws-replicator
no_return_description
Replicate external AWS resources into a running LocalStack instance using the AWS Replicator HTTP API. Returns replication job details for start, job status for status, and the list of jobs or supported resource types for the list and list-resources actions.
localstack-chaos-injector
no_return_description
Injects, manages, and clears chaos faults and network effects in LocalStack to test system resilience. Returns the currently configured fault rules or latency for the get actions, and the result of applying, removing, or clearing faults and latency for the remaining actions.
localstack-cloud-pods
no_return_description
Manages remote LocalStack Cloud Pods: save persists local state to a remote pod, load restores state from a remote pod, and delete removes a remote pod. Returns the result of the performed save, load, or delete action.
localstack-deployer
no_return_description
Deploys or destroys AWS infrastructure on LocalStack using CDK, Terraform, or SAM. Returns the outcome of the deploy/destroy or CloudFormation stack create/delete action, including stack or deployment details and any errors.
localstack-docs
no_return_description
Search the LocalStack documentation to find guides, API references, and configuration details. Returns a list of matching documentation results, up to the requested limit.
localstack-ephemeral-instances
no_return_description
Manage cloud-hosted LocalStack Ephemeral Instances: create, list, fetch logs, and delete. Returns the created or listed instance details, the instance logs for the logs action, or confirmation for delete.
localstack-extensions
no_return_description
Install, uninstall, list, and discover LocalStack Extensions from the marketplace. Returns the list of installed extensions for list, the list of available marketplace extensions for available, and the result of the install or uninstall action.
localstack-iam-policy-analyzer
no_return_description
Configures LocalStack's IAM enforcement and analyzes logs to automatically generate missing IAM policies. Returns the current enforcement mode for get-status, the result of setting the mode for set-mode, and the generated policy details for analyze-policies.
localstack-logs-analysis
no_return_description
Analyzes LocalStack logs to help developers quickly diagnose issues and understand their LocalStack interactions. Returns an analysis summary, error or request breakdowns, or raw log lines depending on the analysisType.
localstack-management
no_return_description
Manage LocalStack lifecycle: start, stop, restart, or check status. Returns the current status for the status action, and the outcome of the start, stop, or restart action.
localstack-snowflake-client
no_return_description
Execute SQL queries and commands against the LocalStack Snowflake emulator using the Snowflake CLI (snow). Use this to run SELECT queries, DDL (CREATE/DROP), DML (INSERT/UPDATE/DELETE), SHOW DATABASES/SCHEMAS/TABLES, DESCRIBE TABLE, and any other Snowflake SQL. Returns the query result rows, execution status, or connection check result.
localstack-state-management
no_return_description
Export, import, reset, and inspect LocalStack state using local file-based workflows on disk. Returns the destination file path and result for export, import/reset confirmation, or the inspected state details for the inspect action.

Selection evidence

Confusable tool pairs.

2 pairs where similar names or overlapping descriptions may send an agent toward the wrong tool.

Tool A Tool B Confidence Why they collide
localstack-management localstack-state-management medium Both names combine 'localstack' and 'management', differing only by 'state'. A task like 'manage my LocalStack state', 'manage the LocalStack instance', or 'handle LocalStack management' could route an agent to the lifecycle tool (start/stop/restart) when the user means export/import/reset of state, or vice versa. The shared token 'management' makes the selection ambiguous without reading the full descriptions.
localstack-cloud-pods localstack-ephemeral-instances low Both manage remote/cloud-backed LocalStack resources with delete capabilities and 'cloud' in scope. A task like 'delete my saved cloud state' or 'manage cloud snapshots' could point at either, and ephemeral-instances even accepts a cloudPod to seed state, blurring the boundary. However, the descriptions are detailed enough that this confusion is only weakly plausible.

Compare the field

One score is useful.
The evidence makes it actionable.

Back to the leaderboard