executing-plans

Use when you have a written implementation plan to execute in a separate session with review checkpoints

Model	Source	Category
opus	core	Workflow

Tools: Read, Glob, Grep, Bash, Edit, Write, Task, Skill

Context: fork

Overview

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

Core principle: Batch execution with checkpoints for architect review.

Mandatory Announcement — FIRST OUTPUT before anything else:

┏━ ⚡ executing-plans ━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ [one-line description of what plan you're executing] ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

No exceptions. Box frame first, then work.

Full Reference

Executing Plans

Overview

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

Core principle: Batch execution with checkpoints for architect review.

Mandatory Announcement — FIRST OUTPUT before anything else:

┏━ ⚡ executing-plans ━━━━━━━━━━━━━━━━━━━━━━━━━━━┓
┃ [one-line description of what plan you're executing] ┃
┗━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━┛

No exceptions. Box frame first, then work.

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

Step 2: Execute Batch

Default batch size: 3 tasks. Adjust based on task complexity:

• Simple tasks (single-file edits, config changes): batch of 4-5
• Medium tasks (new feature with tests): batch of 3
• Complex tasks (multi-file, architectural): batch of 1-2

For each task:

1. Mark as in_progress
2. TDD gate: If the task has implementation steps, verify there is a RED step (failing test) first.
   • If the plan task includes a “Write the failing test” step — execute it, run it, confirm it FAILS before writing any implementation code.
   • If no test step exists in the task, STOP and raise it with your partner before proceeding.
   • REQUIRED SUB-SKILL when unsure how to write the test: Use armadillo:test-driven-development
3. Follow each step exactly (plan has bite-sized steps)
4. Run verifications as specified
5. Mark as completed

Bug encounter between tasks: If a bug or unexpected behavior surfaces during execution:

1. Follow .claude/rules/bug-discipline.md — spawn background subagent
2. Note: ⚠ Bug spawned → <summary> (background)
3. Continue with next task in batch
4. Check bug Task status at end of batch (before Step 3 report)

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 armadillo: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

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

Integration

Required workflow skills:

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