Add PolicyLayer — enforcement layer for AI agents #767

Project: PolicyLayer
GitHub: policylayer/intercept

Description: PolicyLayer is the enforcement layer for AI agents — controlling what they're actually allowed to do. Our open-source proxy, Intercept, sits between agents and their tools (via MCP), enforcing deterministic policies on every tool call before execution.

What it does:

• Enforces spend limits, rate limits, and permissions on agent tool calls
• Deny-by-default — agents only access tools you explicitly allow
• Human-in-the-loop approval for high-risk actions
• Fail-closed — if the enforcement layer is down, nothing executes
• Works with any MCP server (Stripe, GitHub, AWS, Postgres, etc.)

Category suggestion: Infrastructure / Security

Install: npx -y @policylayer/intercept init

License: Apache 2.0