| name | executing-plans |
| description | This skill should be used when the user has a written implementation plan to execute in a separate session with review checkpoints. |
| license | MIT |
Executing Plans
Purpose
Execute approved implementation plans in controlled batches with checkpoints, verification, and review gates.
When to Use
Use this skill when:
- A written implementation plan already exists
- Work must be executed with periodic review checkpoints
- Risk of regressions requires strict plan adherence and verification
Progress Tracking
Source of truth: the plan file’s ## To-do section (checkboxes or ⏳ / ✔️ / ✅).
Also display a short gauge before each batch:
[████░░░░░░░░░░░░░░░░] 20% — Phase 1: Loading & Reviewing Plan
[████████░░░░░░░░░░░░] 40% — Phase 2: Executing Batch N
[████████████░░░░░░░░] 60% — Phase 3: Reporting & Verification
[████████████████░░░░] 80% — Phase 4: Awaiting Feedback
[████████████████████] 100% — Phase 5: Final Verification Complete
For plans with many tasks: X/Y tasks done (N%).
Workflow
- Load and review plan critically
- Execute first batch of tasks
- After each finished task, edit the plan to update To-do
- Report results and verification output
- Pause for feedback and continue in batches
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."
Tools: Agent Pi read / write / edit / bash. Prefer edit for updating the plan’s To-do. Ask the user in chat when blocked.
The Process
Step 1: Load and Review Plan
read the plan file- Review critically - identify any questions or concerns about the plan
- If concerns: Raise them with your human partner before starting
- If no concerns: ensure a
## To-do exists (or treat Suggested order / ⏳ bullets as the To-do), then proceed
Step 2: Execute Batch
Default: First 3 tasks
For each task:
- Mark the To-do item in progress (edit the plan: e.g.
[~] or ⏳ on that line)
- Follow each step exactly (plan has bite-sized steps)
- Run verifications as specified
- After the task is finished:
edit the plan and mark that To-do item done ([x] or ✔️) before starting the next task
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
After all tasks complete and verified:
- Confirm To-do is fully checked off in the plan file
- Summarize what landed and any follow-ups
- Offer optional commit / PR steps if the user wants them
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
- Progress lives in the plan’s To-do — update it with
edit after each finished task
- Between batches: just report and wait
- Stop when blocked, don't guess
- Never start implementation on main/master branch without explicit user consent
Integration
- writing-plans — creates plans this skill executes (with a
## To-do section)
Critical Rules
- Stop immediately on blockers and ask for clarification.
- Run every verification command defined in the plan.
- After each batch, report and wait for feedback before continuing.
- Keep progress in the plan file To-do (
edit after each finished task).
Error Handling
| Error | Likely Cause | Action |
|---|
| Plan file not found | Path provided doesn't exist or file was moved | Ask user to confirm file path; list available plan files in docs/plans/ |
| Ambiguous task description | Step in plan is unclear or underspecified | Stop and ask user for clarification before proceeding |
| Task depends on failed prerequisite | Earlier task failed or was skipped | Report the dependency failure; do not continue until resolved |
| Verification command fails | Tests or checks defined in plan are failing | Stop immediately, report full error output, wait for user instruction |
| Blocker encountered mid-execution | Unexpected state, conflicting files, or missing context | Pause, report the exact blocker, and ask user how to proceed |
| Batch completed with partial failures | Some tasks in batch failed | Report which succeeded and which failed; ask user whether to retry or skip |
| No To-do section | Plan authored without checklist | Add a To-do from task headings (ask first if unclear), then proceed |
Example Usage
- "Use executing-plans to implement
docs/plans/2026-02-20-api-hardening.md."
- "Use executing-plans to run the next 3 tasks and stop for review."
- "Use executing-plans to finish remaining tasks and report final checks."
