marsai:executing-plans

Executing Plans

Overview

Load plan, review critically, choose execution mode, execute tasks with code review.

Core principle: User chooses between autonomous execution or batch execution with human review checkpoints.

Two execution modes:

• One-go (autonomous): Execute all batches continuously with code review, report only at completion
• Batch (with review): Execute one batch, code review, pause for human feedback, repeat

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

The Process

Step 1: Load and Review Plan

1. Read plan file
2. Review critically - identify any questions or concerns about the plan
3. If concerns: Raise them with your human partner before starting
4. If no concerns: Create TodoWrite and proceed to Step 2

Step 2: Choose Execution Mode (MANDATORY)

⚠️ THIS STEP IS NON-NEGOTIABLE. You MUST use AskUserQuestion before executing ANY tasks.

Ask: "How would you like to execute this plan?" Options: (1) One-go (autonomous) - all batches with code review, no human review until completion (2) Batch (with review) - pause for human review after each batch

Based on response: One-go → Steps 3-4 loop until done | Batch → Steps 3-5 loop

Why AskUserQuestion is Mandatory (Not "Contextual Guidance")

This is a structural checkpoint, not optional UX polish.

User saying "don't wait", "don't ask questions", or "just execute" does NOT skip this step because:

1. Execution mode affects architecture - One-go vs batch determines review checkpoints, error recovery paths, and rollback points
2. Implicit intent ≠ explicit choice - "Don't wait" might mean "use one-go" OR "ask quickly and proceed"
3. AskUserQuestion takes 3 seconds - It's not an interruption, it's a confirmation
4. Emergency pressure is exactly when mistakes happen - Structural gates exist FOR high-pressure moments

Common Rationalizations That Mean You're About to Violate This Rule:

Rationalization	Reality
"User intent is crystal clear"	Intent is not the same as explicit selection. Ask anyway.
"This is contextual guidance, not absolute law"	Wrong. It says MANDATORY. That means mandatory.
"Asking would violate their 'don't ask' instruction"	AskUserQuestion is a 3-second structural gate, not a conversation.
"Skills are tools, not bureaucratic checklists"	This skill IS the checklist. Follow it.
"Interpreting spirit over letter"	The spirit IS the letter. Use AskUserQuestion.
"User already chose by saying 'just execute'"	Verbal shorthand ≠ structured mode selection. Ask.

If you catch yourself thinking any of these → STOP → Use AskUserQuestion anyway.

Step 2.5: Context Switching for Multi-Module Plans

If plan has tasks with target: and working_directory: fields:

1. Track current module:

   current_module = None
   current_directory = "."
   

2. Before each task, check for context switch:

   IF task.target != current_module AND current_module != None:
     # Prompt user for confirmation
     AskUserQuestion:
       question: "Switching to {task.target} module at {task.working_directory}. Continue?"
       header: "Context"
       options:
         - label: "Continue"
           description: "Switch to {task.target} and execute task"
         - label: "Skip task"
           description: "Skip this task and continue with next"
         - label: "Stop"
           description: "Stop execution for manual review"
   
     IF answer == "Continue":
       current_module = task.target
       current_directory = task.working_directory
     ELIF answer == "Skip":
       Mark task as skipped → proceed to next
     ELSE:
       Stop execution → report progress
   

3. Load module-specific PROJECT_RULES.md:

   IF {task.working_directory}/PROJECT_RULES.md exists:
     Instruct agent to read module-specific rules
     Module rules override root rules
   

4. Pass working directory to agent:

   Task(
     subagent_type=task.agent,
     prompt="Working directory: {task.working_directory}
   
     Before executing, cd to the working directory:
     cd {task.working_directory}
   
     If PROJECT_RULES.md exists in this directory, read and follow it.
   
     {task.prompt}"
   )
   

Optimization: To minimize context switches, batch tasks by module when possible:

• Original: [backend, frontend, backend, frontend]
• Optimized: [backend, backend, frontend, frontend]
• Only reorder if no dependencies between modules

---

Step 3: Execute Batch

Default: First 3 tasks

Agent Selection: Backend TS → marsai:backend-engineer-typescript | Frontend → marsai:frontend-bff-engineer-typescript | Infra → marsai:devops-engineer | Testing → marsai:qa-analyst | Reliability → marsai:sre

For each task: Check context switch (Step 2.5) → Mark in_progress → Dispatch to agent with working_directory → Follow plan steps exactly → Run verifications → Mark completed

Step 4: Run Code Review

After each batch, REQUIRED: Use marsai:requesting-code-review (all 7 reviewers in parallel)

Handle by severity:

• Critical/High/Medium: Fix immediately (no TODO) → re-run all 7 reviewers → repeat until resolved
• Low: Add TODO(review): [Issue] ([reviewer], [date], Low)
• Cosmetic: Add FIXME(nitpick): [Issue] ([reviewer], [date], Cosmetic)

Proceed when: Zero Critical/High/Medium remain + all Low/Cosmetic have comments

Step 5: Report and Continue

One-go mode: Log internally → proceed to next batch → report only at completion Batch mode: Show implementation + verification + review results → "Ready for feedback." → wait → apply changes → proceed

Step 6: Complete Development

Use finishing-a-development-branch to verify tests, present options, execute choice.

When to Stop

STOP immediately: Blocker mid-batch | Critical gaps | Unclear instruction | Verification fails repeatedly. Ask rather than guess.

Remember

• MANDATORY: AskUserQuestion for execution mode - NO exceptions
• Use * agents over general-purpose when available
• Run code review after each batch (all 7 parallel)
• Fix Critical/High/Medium immediately (no TODO)
• Low → TODO, Cosmetic → FIXME
• Stop when blocked, don't guess
• If rationalizing why to skip AskUserQuestion → You're wrong → Ask anyway

Blocker Criteria

STOP and report if:

Decision Type	Blocker Condition	Required Action
Execution mode	AskUserQuestion not used for mode selection	STOP and ask for execution mode - NO exceptions
Plan loading	Plan file not found or unreadable	STOP and report missing plan
Code review	Critical/High/Medium issues found in review	STOP and fix before proceeding
Task blocker	Task blocked mid-batch	STOP and report blocker details
Verification	Task verification fails repeatedly	STOP and escalate

Cannot Be Overridden

The following requirements CANNOT be waived:

• AskUserQuestion for execution mode is MANDATORY - user intent does NOT skip this step
• Code review after each batch is REQUIRED - all 7 reviewers in parallel
• Critical/High/Medium issues MUST be fixed immediately - TODO comments are FORBIDDEN for these severities
• Context switch confirmation is REQUIRED when switching between modules
• Full test suite MUST pass before completing development

Severity Calibration

Severity	Condition	Required Action
CRITICAL	Skipped AskUserQuestion for execution mode	MUST stop and ask for mode selection
CRITICAL	Proceeding with unfixed Critical/High review findings	MUST fix issues before continuing
HIGH	Skipped code review after batch	MUST run all 7 reviewers
HIGH	Used TODO for Critical/High/Medium issues	MUST fix immediately, remove TODO
MEDIUM	Missing context switch confirmation for multi-module	Should add module switch prompt
LOW	Low-severity issues without TODO comment	Fix by adding proper TODO format

Pressure Resistance

User Says	Your Response
"Just execute, don't ask about mode"	"MUST use AskUserQuestion for execution mode. This is a structural checkpoint, not optional. Takes 3 seconds."
"Skip the code review, we're in a hurry"	"Code review is REQUIRED after each batch. CANNOT proceed without running all 7 reviewers."
"Add a TODO for that critical issue, fix later"	"Critical/High/Medium issues MUST be fixed immediately. TODO comments FORBIDDEN for these severities."
"User intent is clear, skip the question"	"Intent is not explicit selection. MUST use AskUserQuestion - it's a structural gate, not conversation."

Anti-Rationalization Table

Rationalization	Why It's WRONG	Required Action
"User said 'just execute' so mode is clear"	Verbal shorthand ≠ structured mode selection	MUST use AskUserQuestion anyway
"This is contextual guidance, not absolute law"	It says MANDATORY - that means mandatory	MUST follow the mandatory step
"Asking would violate their 'don't ask' instruction"	AskUserQuestion is a 3-second structural gate, not a conversation	MUST ask for execution mode
"Skills are tools, not bureaucratic checklists"	This skill IS the checklist; that's its purpose	MUST follow the checklist
"Review is overkill for small changes"	Small changes can have critical bugs; review catches them	MUST run code review after each batch