A community-driven registry for Claude, Cursor, Windsurf, Cline & more. Not affiliated with Anthropic.
Are you the author? Sign in to claim
An MCP Server that sits between your agent and AWS STS and issues temporary credentials scoped to specific AWS Services
(or any AI agent)
An MCP server that sits between your AI agent and AWS STS, issuing temporary credentials scoped to specific AWS services and access levels on demand.
Timebound-IAM is an MCP Server that issues short-lived, service-scoped AWS credentials via STS AssumeRole so that AI coding agents (like Claude Code) can access AWS resources without long-lived keys. Credentials are time-bounded (15 minutes to 12 hours), scoped to specific services and access levels (read-only or full), and automatically cleaned up on expiry.
Homebrew (macOS/Linux)
brew install builder-magic/tap/timebound-iam
Go install
go install github.com/builder-magic/timebound-iam@latest
Binary download — Download pre-built binaries from GitHub Releases.
For the complete installation and setup guide, see https://timebound-iam.com/installation-and-setup.
Configure AWS
Run the setup wizard to generate the IAM trust policy and inline policy for the broker role:
bin/timebound-iam setup aws
# or specify a named profile
bin/timebound-iam setup aws --profile my-profile
Follow the printed instructions to create the timebound-iam-broker IAM role in your account with the generated policies.
Add to Claude Code
Register the MCP server so Claude Code can request temporary credentials on demand:
claude mcp add --scope user timebound-iam -- timebound-iam serve
Restart Claude Code to pick up the new server.
Verify
Verify that the MCP server is installed and running with the /mcp command:
Test the credential flow end-to-end:
timebound-iam test
This requests short-lived S3 read-only credentials and writes them to a temporary .env file you can use to verify access.
Timebound-IAM can also be used as a standalone CLI to wrap shell scripts with scoped, temporary credentials.
exec
Runs a command with temporary credentials injected into its environment:
timebound-iam exec -s s3:full,cloudfront:full -t 30m --profile prod -- ./deploy.sh
Use --dry-run to validate flags without requesting credentials. Use --no-confirm to skip the interactive prompt in CI or scripts.
env
Prints export or unset statements for use with eval:
eval "$(timebound-iam env -s s3:ro -t 15m --no-confirm)"
aws s3 ls
eval "$(timebound-iam env --unset)"
Scopes
Scopes use the format service:level where level is ro (read-only) or full. Multiple scopes can be comma-separated or passed as repeated -s flags:
-s s3:ro,dynamodb:full
-s s3:ro -s dynamodb:full
Contributions in any form (suggestions, bug reports, pull requests, and feedback) are welcome. If you've found a bug, you can submit an issue or email me at rsingh@builder-magic.com.
This project is licensed under the Apache License, Version 2.0.
Unless you explicitly state otherwise, any contribution intentionally submitted for inclusion in the work by you, as defined in the Apache-2.0 license, shall be licensed under the Apache License, Version 2.0, without any additional terms or conditions.
Contact: rsingh@builder-magic.com
MCP server integration for DaVinci Resolve Studio
A trilingual (繁中 / English / 简中) learning roadmap for agentic AI: from LLM basics to multi-agent systems, with 240+ cura
Run Claude Code as an MCP server so any agent can delegate coding tasks to it
Browser automation using accessibility snapshots instead of screenshots