Skip to content

Playbooks

Playbooks show how sivtr works in practice: real scenarios where an agent uses local workspace memory.

These short recordings show the core loop: capture local work, search it, narrow the evidence, and hand exact context to the next command or agent.

Search recent terminal output with sivtr
Find the latest terminal evidence.
Browse and reuse captured workspace memory
Browse and reuse captured context.
Build a timeline from local agent and terminal memory
Turn recent work into a timeline.
Pass a named memory variable through a command chain
Save matches as variables and keep working.
PlaybookWhat it demonstrates
Fix the latest terminal errorThe agent finds the failure, patches the issue, and verifies the fix.
Build a recent work timelineThe agent reconstructs what you worked on from timestamps, commands, and agent dialogues.
Continue after interruptionThe agent searches memory to find the most recent work thread.
Agent handoffThe agent prepares a structured handoff with evidence and next steps.
Remote collaboration memoryFuture direction: search a teammate’s agent memory with permission.