Manage outcomes and task tracking
docommandsetup L1★124
one-ie/one ↗What it does
Drive work through intent or TODO files
Best for
Managing multi-wave work streams where wave-dependent model routing and context-aware task selection accelerate delivery over manual task picking.
Inputs
- · natural language intent string OR TODO filename OR --auto/--wave flags
Outputs
- · reflection block showing topic/context/decision
- · execution of selected task or wave
- · outcome marked via /signal
Requires
- · TypeDB query interface
- · project TODO files
- · /signal skill
- · language model routing (Haiku/Opus/Sonnet per wave)
Preconditions
- · docs/TODO.md exists and loaded
- · task-wave attributes set (W1/W2/W3/W4)
- · git repository initialized
Failure modes
- · TODO file not found (ask user for scope)
- · task-wave attribute missing (fallback to EFFORT_MODEL)
- · greenfield task with no existing context (create minimal TODO)
- · skill pre-flight check fails (block execution)
Trust signals
- · wave-aware model routing (Haiku→Opus→Sonnet matching task complexity)
- · skill pre-flight check before execution
- · reflection block forces decision transparency
- · signal outcomes after execution