For Platform and Security

Use this page if you are evaluating AI coding tools through the lens of permissions, credentials, MCP risk, review controls, and revocation.

Who This Is For

• platform engineering teams
• security engineers
• compliance-minded engineering leads
• internal developer platform maintainers

Best First Steps

1. Read Security and Permissions.
2. Review What is MCP?.
3. Apply MCP Approval Policy.
4. Use Security Checklist Template during review.

Highest-Priority Pages

Permissions and Approval

• Security and Permissions
• MCP Approval Policy
• Approved Tools Policy

MCP Risk and Evaluation

• What is MCP?
• Useful MCP Servers
• GitHub MCP
• Filesystem MCP
• Supabase / Postgres MCP

Review Workflows

• Security Review
• Code Review
• Prompt Review Policy

Reusable Controls

• Tool Evaluation Template
• Security Checklist Template
• Customize for Your Team

Recommended Next Action

Start with a read-only tool or MCP pilot, define revocation before broader approval, and keep merge or deploy rights behind explicit human review.

Verification Note

This page points to the strongest security-facing material in the repo. Some MCP pages remain evaluation guides because implementation-level risk varies.