• Jono Nourse
• 07 Oct, 2026
• • ai-agents
  • agent-permissions
  • risk-based-permissions
  • least-privilege
  • human-approval
  • agent-governance
  • agent-security
  • autohive

The Case for Risk-Based Agent Permissions

Approving every agent action leads to rubber-stamping, and treating drafts as safe misses real risk. This post shows how to set permissions by worst …