Evaluate and redact sensitive AI content locally before optional Bedrock Guardrails and Presidio controls, with fail-closed policy decisions and metadata-only evidence.
-
Updated
Nov 2, 2026 - Python
Evaluate and redact sensitive AI content locally before optional Bedrock Guardrails and Presidio controls, with fail-closed policy decisions and metadata-only evidence.
This is a process to install Openshift 4 in a disconnected way
Open source, self-hosted MCP and model gateway: approved models, MCP Tools, budgets, and permissions for your people and coding agents, in your own cloud. MIT, community maintained.
Plan and govern bounded AWS FIS and extension experiments with GovCloud-aware validation, rollback contracts, and reviewable evidence.
Looks valuable. Leaves evidence. Offline cyber decoy detection and assurance for AWS GovCloud and Windows environments.
Offline-capable AI copilot for forward operating bases: YOLOv8 drone detection, multi-sensor Kalman fusion, LangGraph multi-agent triage, and human-in-the-loop approval — runs fully without cloud connectivity (DDIL), syncs to AWS GovCloud on link recovery.
Surface potential waste, security blind spots, and operational risk across 29 AWS GovCloud services using one Python file and one command.
Terraform reference architecture for federal LLM workloads: no-egress VPC, KMS everywhere, NIST 800-53 mapping
Opinionated Terraform modules for a secure AWS platform baseline
Terraform module: ByMorning, an open source, self-hosted MCP and model gateway, on AWS and AWS GovCloud (mirror of installers/aws/modules/bymorning)
To associate your repository with the aws-govcloud topic, visit your repo's landing page and select "manage topics."