executing-plans

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.

Announce at start: "I'm using the 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 user before starting
4. If no concerns: Create tasks via TaskCreate 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. Record discoveries — After each task, append unexpected findings, decisions, or technical insights to .writing/findings.md, one entry per finding:

   ## [Task N] <short title>
   - **What:** <the discovery/decision in one line>
   - **Why it matters:** <impact on later tasks or the plan>
   - **Action:** <kept / changed approach / needs user input>
   

   The schema is a format, not a scope filter — still record decisions, surprises, and consolidated batch-level insights (Step 3), not only per-task findings.
5. Mark as completed

Step 3: Report and Update Progress

When batch complete:

• Show what was implemented
• Show verification output
• Update .writing/progress.md (if .writing/ exists):
  • Mark completed tasks as complete in the Task Status Dashboard
  • Append a batch summary to the session log section: Batch <n>: tasks <ids> complete | verified: <command/result> | open: <blockers or none>
• Update .writing/findings.md — Consolidate any discoveries, decisions, or surprises from this batch
• 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:

• Read .writing/progress.md to compile a full summary of all batches and verification evidence.
• Run superpower-writing:claim-verification if the manuscript is ready for an evidence-reliability pass.
• Present a concise completion summary: what was drafted, claim/citation status, any remaining [NEEDS-EVIDENCE] or % draft-only markers, and the recommended next step (hand the skeleton to the human author for refinement, or archive via superpower-writing:archiving).
• Do NOT merge, create a PR, or run destructive git cleanup unless the user explicitly asks in this session.

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
• After each task, record discoveries to .writing/findings.md
• After each batch, update both .writing/progress.md and .writing/findings.md
• Before final report, read .writing/progress.md for full summary

Integration

Related skills:

• superpower-writing:git-worktrees - recommended when starting from a shared or risky branch
• superpower-writing:writing-plans - creates the plan this skill executes
• superpower-writing:claim-verification - evidence-reliability pass once sections are drafted
• superpower-writing:archiving - archive the completed .writing/ state when done