# Environment Aliases Identify Targets, Not Authority An environment alias identifies where a command runs, not whether the command is permitted. ## Scope & conditions - This applies when a tool uses a short name such as `@prod` to route commands to a remote system. - The alias belongs to the connection and naming layer. Authorization still needs a separate policy layer that defines read, write, approval, and destructive-command rules. - The distinction matters most for agents because a convenient target name makes production access easier without making production changes safer. ## Source anchor > "Production access is read-only unless I explicitly approve a write." <!-- atomicity: 1 idea ✓ --> ## Links - [[3. Resources/Permanent Notes/Bounded Action Spaces Not Better Models Make Agents Safe to Run]] — applies-in: a named target still needs a bounded set of allowed actions. - [[3. Resources/Permanent Notes/Safe Agent Changes Come From Guardrails and Observability Not Trust]] — extends: approval rules and audit trails govern what may happen after an alias selects the target. - [[AI Agent Safety MOC]] — belongs-in: target selection and command authority are separate safety controls. ## Source - [[3. Resources/Readwise/Articles/How WP-CLI Aliases Supercharge Your Agents - Brian Coords]] — separates SSH connection, WP-CLI environment names, and project-level permission rules.