forge-agent

To isolate task context and prevent main agent performance bottlenecks, you can spawn a fresh sub-agent to execute tasks separately. This keeps the primary context clear while the sub-agent handles isolated execution.

Fresh context analysis evaluates tasks without prior conversation biases by using a newly spawned sub-agent. The sub-agent runs the analysis in an isolated environment, ensuring objective blind testing results.

You can automate forge tasks by triggering a command that automatically spawns a general-purpose sub-agent. This sub-agent executes the specified forge tasks independently, maintaining the main context's performance.

Yes, you need the Agent tool and forge skill framework to run isolated sub-agent tasks. These dependencies are required to automatically spawn the sub-agent and execute the forge tasks in an isolated context.

You should use context isolation for blind testing when you need to evaluate task outputs without prior context contamination. Spawning a fresh sub-agent ensures the test environment remains completely isolated from the main context.