Skip to content

majiayu000/spellbook / skills/multi-model-orchestrator

multi-model-orchestrator

Verified

Use when coordinating complex tasks across multiple AI agents with a centralized handoff document for planning, execution tracking, and feedback fusion.

templates/handoff-template.yaml

View on GitHub
---
# Handoff Template for Multi-Model Orchestration
# Copy this file and fill in your own goal and subtasks

metadata:
  name: "my-awesome-task"
  created: "2026-06-12"
  goal: "What are we trying to accomplish?"
  status: "planning"

goal:
  summary: |
    Clear, specific description of what we're trying to accomplish.
    What problem does this solve? Who benefits?

  acceptance_criteria:
    - "Testable criterion 1"
    - "Testable criterion 2"
    - "Testable criterion 3"

  context: |
    Any background needed to understand the task:
    - Prior decisions
    - Constraints
    - Related code or documentation
    - Assumptions

subtasks:
  # Example subtask structure
  - id: "task-1"
    title: "Analyze and design"
    type: "analysis"  # Options: code | analysis | design | research | validation
    executor: "Opus"  # Who will execute: Opus | Haiku | Codex | Claude | Claude Code

    input: |
      What does the executor need to know?
      Include:
      - Specific problem statement
      - Code references (file paths, line numbers)
      - Examples of current behavior
      - Constraints or requirements
      - Links to related documentation

    acceptance: |
      How do we know this subtask is complete?
      Be specific and testable:
      - Design document created
      - All 5 edge cases covered
      - Performance < 100ms

    depends_on: []  # List of task IDs this depends on (empty = can start immediately)
    status: "pending"

  - id: "task-2"
    title: "Implement the solution"
    type: "code"
    executor: "Codex"

    input: |
      Based on the design from task-1:
      - Implement in src/module.ts
      - Follow existing patterns in src/similar.ts
      - Add tests in tests/module.test.ts

    acceptance: |
      - All tests pass
      - Code review standards met
      - No performance regression

    depends_on: ["task-1"]  # Wait for task-1 to complete
    status: "pending"

  - id: "task-3"
    title: "Test and validate"
    type: "validation"
    executor: "Claude Code"  # Interactive execution

    input: |
      Run the implementation end-to-end:
      1. Start the dev server
      2. Execute the test scenarios
      3. Verify performance
      4. Check edge cases

    acceptance: |
      - All tests pass
      - Performance acceptable
      - No regressions in related features

    depends_on: ["task-2"]
    status: "pending"

execution:
  rounds: []
  # This will be populated as you execute each subtask
  # Example entry:
  # - round: 1
  #   task_id: "task-1"
  #   executor: "Opus"
  #   status: "done"
  #   result: |
  #     Design document created at docs/design.md
  #     Key insights:
  #     - Strategy A is fastest
  #     - Need mutex for race conditions
  #   issues: |
  #     Edge case with concurrent requests still unclear
  #   next_step: "Clarify edge case, then proceed to task-2"
  #   timestamp: "2026-06-12T10:30:00Z"

feedback:
  synthesis: |
    (Fill in after execution)
    Overall assessment: what worked? what didn't?

  improvements:
    - "Potential improvement 1"
    - "Potential improvement 2"

  decision_log: |
    Why did we make certain choices?
    What tradeoffs were involved?
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 472kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall