Read MPI task cards and plans to propose next actions.
npx skills add https://github.com/MadPonyInteractive/mpi-kanban --skill mpi-continue
Please help me install this Agent Skill. Skill: mpi-continue Source: https://github.com/MadPonyInteractive/mpi-kanban/tree/main/skills/mpi-continue Command: npx skills add https://github.com/MadPonyInteractive/mpi-kanban --skill mpi-continue
Continue active MPI work intelligently by reading the current task card, the active plan, and the coordination state to determine the next best action. It replaces rigid "execute next" behavior with context-aware progression and bounded, safe card updates.
Tell me to continue the MPI plan and I will read the current card and plan, then propose the next action.
High-intent search queries and answers about installing and using this skill.
To continue work, you read the current task card, active plan, and coordination state to determine the next best action. The system proposes concrete next steps and updates task state only after user approval.
Handoff-aware task coordination uses a shared coordination state to synchronize work across agents. It reads this state alongside the active plan to intelligently progress tasks and ensure safe, context-aware transitions.
Yes, you can inspect a specific JSON task card without changing its state using a read-only board entry mode. This allows you to surface the current card state and review details before proposing any updates.
The best way to safely update task states is using direct card update mode with safety checks and in-bound transitions. It ensures state mutations are bounded, context-aware, and only execute after user approval.
Yes, you need plan derivations to determine the next action. The system derives concrete next steps from the active MPI plan, requiring access to board files, plan derivations, and coordination references to function correctly.
MadPonyInteractiveClose active MPI sessions, sync rules/docs, and commit touched files.
tody-agentRead task.md and output a concise status with completed task, next steps, and blockers.
abdullah1854Coordinate Kanban boards with RAG-driven knowledge retrieval and task sync.
anorthamRestore session context from persistent memory at coding session start.
