Nine source-backed chapters on bounded scouts, operating homes, semantic state, merge receipts, cost measurement and safe adoption.
Latest article
Trace task state, semantic busy verdicts and generation-bound wake handling without confusing a visible terminal with a completed or recoverable task.
By publication date
01 → 09
Trace task state, semantic busy verdicts and generation-bound wake handling without confusing a visible terminal with a completed or recoverable task.
Select by task independence, human decision points and durable-state requirements rather than treating multi-agent execution as a universal upgrade.
Plan a local agent operating environment rather than a web-server deployment, with separate private state, deliberate credentials and versioned updates.
Prepare an isolated trial, distinguish inspection commands from launching trusted hooks, and define a read-only investigation before permitting changes.
Design a learning project that displays task identities, observation sources and merge receipts without turning a status view into an unauthorized control plane.
Understand the agent distro, its ship and scout tasks, and the difference between terminal visibility, durable state and permission to merge.
Interpret zero-token watcher supervision narrowly and design a useful trial that measures accepted work, model usage and human recovery effort.
Separate workflow policy from a security sandbox, preserve explicit merge permission and understand documented race limits before unattended operation.
Follow authority resolution, six-line persistence and identity-checked consumption, including the limits of an isolated JavaScript teaching model.
