Published work filed under MCP Security.
After the OpenCode breach, revoke AI API keys at the provider, set hard spend limits, use one key per tool, and keep secrets out of MCP config files and agents.
How to apply least privilege to your agent's MCP tools: default-deny allowlists, narrow scopes, per-server credentials, approval for writes, and a runtime gate.
This checklist shows how to vet an MCP server before an agent touches it, from provenance and reading a published grade honestly to auth, version pinning and cr…
Point-in-time MCP grades find risk; runtime firewalls enforce policy on tool calls. Andrax mcpgrade-1.4.0 vs Guard mcpgrade-2.0.0 — no score equating.
