[50] EXECUTE. Execute plans step-by-step with confirmation gates. Each step requires user approval before proceeding. Includes change management lifecycle (Pre-Change → During → Post-Change → Rollback). Use when implementing approved plans, deploying changes, or any multi-step execution requiring control and reversibility.
Core Principle: Execute with control. Every step gets a gate — confirm before proceeding. Build in rollback points. Maintain audit trail.
When you invoke this skill, you're asking AI to:
| Dimension | 100% Core | +50% Enhancement |
|---|---|---|
| Steps | Execute sequence | + Atomic, reversible actions |
| Gates | Get confirmation | + Offer Continue/Pause/Research/Replan |
| Documentation | Log actions | + Full audit trail |
| Rollback | Have revert plan | + Pre-identified safe points |
Declare execution mode and current state:
⚙️ **Gated-Exec 150 Activated**
**Plan:** [What we're executing]
**Total Steps:** [N]
**Current Phase:** [1 of N]
**Rollback Points:** [Where we can safely stop]
Before each step:
Single atomic step:
After each step:
Present decision point:
✅ **Step [X/N] Complete**
**Action:** [What was done]
**Result:** [Outcome]
**Next:** [What's coming]
**Options:**
- **Continue** → Proceed to next step
- **Pause** → Stop for clarification
- **Research** → Need more investigation
- **Replan** → Strategy needs modification
- **Rollback** → Revert to safe point
🔄 CHANGE LIFECYCLE
PRE-CHANGE (Before each step)
├── Impact analysis complete
├── Risk assessment done
├── Backup strategy ready
└── Rollback plan documented
DURING CHANGE (Execution)
├── Atomic step execution
├── Real-time monitoring
├── Immediate issue detection
└── Progress tracking
POST-CHANGE (After each step)
├── Outcome verification
├── Integration testing
├── Performance validation
└── Documentation update
ROLLBACK READY (Always)
├── Safe reversal path
├── State restoration
├── Impact minimization
└── Lesson documentation
| Situation | Action | Reason |
|---|---|---|
| Step Success | Get confirmation, proceed | Maintain control |
| Minor Issue | Document, continue with caution | Don't stop for small problems |
| Major Problem | Pause, investigate, possibly rollback | Prevent cascading failures |
| Plan Deviation | Replan with user approval | Maintain strategic alignment |
| New Information | Research, update plan | Adapt to new insights |
During execution:
⚙️ **Gated-Exec: Step [X/N]**
**Phase:** [Current phase name]
**Action:** [What we're doing]
[Execution details...]
**Result:** ✅ Success | ⚠️ Issues | ❌ Failed
**Status:**
├── ✅ Completed: [Steps done]
├── 🔄 Current: [This step]
└── ⏳ Pending: [Steps remaining]
**Rollback Point:** [Last safe state]
**Continue?** (Yes / Pause / Research / Replan / Rollback)
| Failure | Detection | Recovery |
|---|---|---|
| Phase Skipping | No confirmation obtained | Return to skipped gate |
| Poor Documentation | Missing audit trail | Document current state, continue with logging |
| Change Cascades | Unexpected side effects | Rollback to safe point, investigate |
| Loss of Control | Process confusion | Emergency pause, status assessment |
User: "Deploy the new feature"
AI: [Deploys all 5 steps without confirmation]
Result: Breaking change deployed, emergency rollback needed
User: "Deploy the new feature"
⚙️ Gated-Exec 150 Activated
Plan: Feature deployment
Total Steps: 5
Rollback Points: After step 2, After step 4
---
⚙️ Gated-Exec: Step 1/5
Phase: Database Migration
Action: Run migration script
[Migration executed successfully]
Result: ✅ Success
- 3 tables updated
- No data loss
- Backup created
Continue? (Yes/Pause/Research/Replan/Rollback)
User: "Yes"
⚙️ Gated-Exec: Step 2/5
...
After completing this skill, write to .sessions/SESSION_[date]-[name].md:
### [date - HH"MM] Execute-Gated 150 Complete
**Plan Executed:** <plan name>
**Steps Completed:** <X/N>
**Outcome:** <Success/Rolled Back>
**Artifacts:** <files changed>
Remember: Gated execution isn't slow — it's safe. The gates prevent costly mistakes and keep you in control. Every "Yes" is a conscious decision, every step is reversible.