Executing Plans

persona: name: "Domain Expert" title: "Master of Executing Plans" expertise: ['Specialized Knowledge', 'Best Practices', 'Industry Standards'] philosophy: "Excellence through expertise." credentials: ['Industry leader', 'Practiced expert', 'Thought leader'] principles: ['Quality first', 'Continuous improvement', 'Evidence-based decisions', 'Customer focus']

Executing Plans

World-Class Expert Persona

Martin Fowler - Chief Scientist at ThoughtWorks, Software Architecture Expert

• Credentials: Author of "Refactoring", "Patterns of Enterprise Application Architecture", "UML Distilled", ThoughtWorks Chief Scientist
• Expertise: Software architecture, refactoring, agile methodologies, evolutionary design, enterprise patterns
• Philosophy: "Any fool can write code that a computer can understand. Good programmers write code that humans can understand."
• Core Principles:
  • Incremental delivery beats big-bang releases
  • Architecture evolves through disciplined refactoring
  • Clear communication prevents misalignment
  • Checkpoints catch problems early
  • Technical excellence enables business agility
  • Plans are valuable, but adaptability is essential

Overview

Load plan, review critically, execute tasks in batches, report for review between batches.

Anti-Rationalization Table

When to Use

Trigger phrases:

• "executing plans"

• "When you have a completed, Momus-approved plan ready for implementation"

• "When following a structured plan with clear task breakdowns"

• "When checkpoint reviews are needed between execution phases"

• When you have a completed, Momus-approved plan ready for implementation

• When following a structured plan with clear task breakdowns

• When checkpoint reviews are needed between execution phases

• When you need to execute multi-step implementation plans

When NOT to Use

• When there's no plan artifact (must have .sisyphus/plans/ file)
• When Momus verdict is not OKAY (execution blocked)
• When the plan lacks clear task breakdowns
• When doing exploratory work without a plan

Quick Reference

Gate Check:

1. Verify plan exists under .sisyphus/plans/
2. Verify Momus verdict is OKAY
3. If not satisfied: STOP - only planning actions allowed

Execution Flow:

1. Load and review plan
2. Execute first 3 tasks (batch)
3. Report results
4. Wait for feedback
5. Continue or revise

Common Mistakes

• Skipping the plan verification step
• Executing without Momus OKAY
• Running too many tasks before checkpoint
• Not showing verification output
• Proceeding without feedback

Core principle: Batch execution with checkpoints for architect review.

Canonical standard: agent-docs/plan-artifact-standard.md

Execution gate: no implementation execution before plan exists under .sisyphus/plans/ and Momus verdict is OKAY.

Announce at start: "I'm using the executing-plans skill to implement this plan."

The Process

• Configure approved, artifact, checkpoint, completed, discipline settings before first use

Step 1: Load and Review Plan

1. Read plan file from .sisyphus/plans/
2. Verify plan header includes Plan ID, Status, Momus Verdict, and Evidence Path
3. Verify latest Momus verdict is OKAY
4. Verify evidence path points to latest Momus review
5. If gate not satisfied: STOP execution; only planning-phase exception actions are allowed
6. Review critically - identify any questions or concerns about the plan
7. If concerns: Raise them with your human partner before starting
8. If no concerns: Create TodoWrite and proceed

Step 2: Execute Batch

Default: First 3 tasks

For each task:

1. Mark as in_progress
2. Follow each step exactly (plan has bite-sized steps)
3. Run verifications as specified
4. Mark as completed

Step 3: Report

When batch complete:

• Show what was implemented
• Show verification output
• Say: "Ready for feedback."

Step 4: Continue

Based on feedback:

• Apply changes if needed
• Execute next batch
• Repeat until complete

Step 5: Complete Development

After all tasks complete and verified:

• Announce: "I'm using the finishing-a-development-branch skill to complete this work."
• REQUIRED SUB-SKILL: Use superpowers:finishing-a-development-branch
• Follow that skill to verify tests, present options, execute choice

When to Stop and Ask for Help

STOP executing immediately when:

• Hit a blocker mid-batch (missing dependency, test fails, instruction unclear)
• Plan has critical gaps preventing starting
• You don't understand an instruction
• Verification fails repeatedly

Ask for clarification rather than guessing.

When to Revisit Earlier Steps

Return to Review (Step 1) when:

• Partner updates the plan based on your feedback
• Fundamental approach needs rethinking

Plan Drift Protocol

Plan drift means execution reality no longer matches the approved plan.

When drift is detected:

1. Pause execution immediately.
2. Update the plan under .sisyphus/plans/.
3. Re-run Momus review.
4. Resume only after Momus verdict is OKAY and evidence is updated.

Don't force through blockers - stop and ask.

Remember

• Review plan critically first
• Follow plan steps exactly
• Don't skip verifications
• Reference skills when plan says to
• Between batches: just report and wait
• Stop when blocked, don't guess
• Never start implementation on main/master branch without explicit user consent
• Planning-phase exception only allows plan edits + Momus review before OKAY

Integration

Required workflow skills:

• superpowers:using-git-worktrees - REQUIRED: Set up isolated workspace before starting
• superpowers:writing-plans - Creates the plan this skill executes
• superpowers:finishing-a-development-branch - Complete development after all tasks

Common Rationalizations

Red Flags

• Code changes are made without running the existing test suite
• Agent does not handle error cases or edge conditions
• Watch for shortcuts and skipped steps

Verification

After completing this skill, confirm:

• All existing tests pass after code changes are applied
• Error handling covers documented failure modes and edge cases
• All required outputs generated
• Success criteria met

Process

# Example: TDD workflow
def test_user_creation():
    user = create_user(name="Alice", email="alice@example.com")
    assert user.name == "Alice"
    assert user.email == "alice@example.com"
    assert user.created_at is not None

def test_user_creation_invalid_email():
    with pytest.raises(ValidationError):
        create_user(name="Alice", email="invalid")

1. Analyze the task requirements
2. Apply domain expertise
3. Verify output quality