Case Lifecycle
Work tracking and progression. How work flows from idea to completion.
Vision
How work flows from idea to completion. The progression model, decision points, the human-in-the-loop philosophy.
Implementation
Automated skill that drives cases through their lifecycle stages.
Structured presentation of inputs with decision prompts for human review.
Autonomous evaluation of inputs — the system decides on routine items without human intervention.
Algorithmic scoring to rank cases by priority and importance.
Assessment of input complexity to determine routing and handling approach.
Inputs inherit priority from their parent case for consistent handling.
Scheduled hourly job that runs case progression across all active cases.
Notes & Future Ideas
Staleness patrol — periodic sweep of all open cases and inputs to detect drift. Flags items as superseded (another case/change already covers this), context-drifted (original framing outdated due to system or life changes), or dead (no longer relevant). Importance-aware thresholds — high-priority items flag sooner. Suggests close, rewrite, or merge. Never auto-closes — always surfaces for human decision.
Parallel work streams — cases can have multiple active tasks running simultaneously, not just one sequential input at a time. Each task carries a mode: collaborative (requires a live session with the human — discussion, review, creative work) or autonomous (the system executes independently and reports back). Default remains sequential for simple cases. Parallel mode is opt-in when a case naturally has independent threads — e.g. research that can run in the background while a separate decision waits for human input. The system proposes parallelization when it detects independent work; the human approves the split. Collaborative tasks queue for live sessions. Autonomous tasks get dispatched, executed, and reported asynchronously.