Review aids you complete locally

Free, with no signup. Editable records work offline; worked examples are fictional. The operations kit also includes an executable lab that creates and cleans up its own synthetic workloads.

Before the agent acts on what it readWhich boundary decides: approval, release, memory, tool, retrieval?Five editable records with proposed acceptance cases: what you approved, what may leave the context, what memory may instruct, what changed in a tool, and what retrieval does not grant.
Before you delegate the jobMake the delegation inspectableA complete fictional research dossier and a transparent workflow cost model that pair with the six-article delegation series.
When the run says doneFinish the task. Account for what remains.An executable reference lab and editable records for run receipts, session ownership, budget contracts, plugin authority and incident handoffs.
Before an unattended runWhat can this agent actually change?A field guide, editable record and 24 proposed scenarios covering authority, process ownership, budgets, evidence and changing tools.
Before you enable accessBefore you connect an agent to a private repositoryA worked parser-fix review and a blank decision record: the access the task needs, the boundary that enforces it, and the evidence still missing.
After a failed writeThe tool timed out. Did the action happen?A fictional timeline and local record that separate the intended operation, each attempt and the external effect before anyone retries.