SuperAPI AI Directory

shrimp-task-manager

Agents & models

View repository

Updated 10-16 10:22

About this skill

Shrimp Task Manager

Quick start

Goal: turn a large task into an explicit task system with analysis, decomposition, execution guidance, and verification.

Reference notes: skills/shrimp-task-manager/references/shrimp-task-manager.md

Best-fit use cases

  • Multi-step implementation with dependencies.
  • Research-heavy engineering work.
  • Tasks needing explicit acceptance criteria and progress tracking.
  • Cases where you want a scored verification step before claiming done.

Recommended workflow

  1. Analyze the task: analyze_task or plan_task
  2. Break it down: split_tasks
  3. Inspect and execute one task at a time: list_tasks, get_task_detail, execute_task
  4. Update tasks as reality changes: update_task
  5. Verify completion: verify_task

Important behavior

  • execute_task gives guidance; it does not perform the work for you.
  • verify_task should reflect the real outcome against the recorded criteria.
  • Use research_mode when a technical investigation spans multiple passes.

Guardrails

  • Do not fabricate requirements; gather evidence first.
  • Keep task granularity reasonable and dependencies explicit.
  • Avoid deleting unfinished tasks unless there is a clear reason.

References

  • Condensed notes: skills/shrimp-task-manager/references/shrimp-task-manager.md