The open firewall for AI agents.
This repository hosts the organization profile and community health files for the Atlas Shield GitHub organization.

Atlas Shield publishes open specifications, SDKs, and reference implementations to protect LLM/AI agents from prompt injection, data exfiltration, and unsafe tool/OS actions. Everything is MCP-compatible and open source where possible.

• Maintainers: oss@atlasshield.ai
• Security (responsible disclosure): security@atlasshield.ai

We welcome issues and PRs!

• For major changes, please open an Issue first to discuss scope.
• Use feature branches (feat/...) and Conventional Commits.
• Do not report vulnerabilities publicly—email security@atlasshield.ai instead.

See also: CONTRIBUTING.md and CODE_OF_CONDUCT.md (in this repo).

• Org profile: profile/README.md (shown on the organization homepage)
• Policies: SECURITY.md, CONTRIBUTING.md, CODE_OF_CONDUCT.md
• Templates: .github/ISSUE_TEMPLATE/*, .github/PULL_REQUEST_TEMPLATE.md
• Specs & policies: the public repo atlas-shield-spec (once live)

We keep it minimalist: safety first, features second. 😉

ℹ️ This .github repository contains no product code. Licenses live in individual code repositories (default Apache-2.0).