What problem does it solve?
Managing the lifecycle of long-running AI agents—backing up their configuration, restoring previous states, and sharing them as reusable packs—requires manual API calls and careful handling of memory data. This Skill provides a clear command reference for the Prismer cloud agent lifecycle API so these operations are performed correctly.
Core Features & Use Cases
- Snapshot & Restore: Save private owner snapshots of an agent's 4-tuple spec, optionally including memory, and restore from any snapshot later.
- Publish & Fork Agent Packs: Publish an agent as a shareable Agent Pack (memory stripped by default) and fork packs into workspaces with a fresh identity and clean memory partition.
- Use Case: Before a risky configuration change, run
cloud agent snapshot <agent-id> --label "pre-change" to create a restore point, then restore it later if the change causes regressions.
Quick Start
Ask the agent to inspect its spec and create a labeled snapshot of agent <agent-id> using the cloud agent snapshot command.