Plan-Action 150 Protocol

Core Principle: Know the path before walking. Have clear steps, know what success looks like, and be ready for what could go wrong.

What This Skill Does

When you invoke this skill, you're creating:

• Clear Steps — What to do and in what order
• Success Criteria — How to know when done
• Risk Awareness — What could go wrong and how to handle it
• Fallback Options — Safe points and alternatives

The Plan Structure

📋 ACTION PLAN
├── 🎯 Goal: What we're achieving
├── 📝 Steps: Ordered actions
├── ✅ Success: How we know we're done
├── ⚠️ Risks: What could go wrong
└── 🔄 Fallback: How to recover

The 150% Planning Rule

• 100% Core: Clear steps + success criteria
• 50% Enhancement: Risks identified + fallback options ready

When to Use This Skill

Universal trigger: Before doing anything with multiple steps or significant complexity.

Specific triggers:

• Starting a new feature or project
• Making changes to existing systems
• Deploying or releasing
• Refactoring or migrations
• Any work where "just do it" isn't enough
• When asked "how should we approach this?"

Key insight: Good plans prevent chaos. Time spent planning saves time spent fixing.

Execution Protocol

Step 1: DEFINE THE GOAL

Be crystal clear on what success looks like:

• What exactly are we trying to achieve?
• What does "done" look like?
• What's in scope, what's out?

Step 2: BREAK INTO STEPS

Decompose into actionable pieces:

• What are the logical phases?
• What's the order (what depends on what)?
• What are the atomic actions?

Step 3: SET SUCCESS CRITERIA

Define how to measure completion:

• How do we know each step is done?
• What tests/checks confirm success?
• What's the quality bar?

Step 4: IDENTIFY RISKS

Think about what could go wrong:

• What are the failure points?
• What has highest impact if it fails?
• What's most likely to go wrong?

Step 5: PLAN FALLBACKS

Prepare for problems:

• Where are the safe stopping points?
• How do we rollback if needed?
• What's plan B?

Step 6: ESTIMATE & VALIDATE

Reality check the plan:

• Is this realistic given resources?
• Are dependencies available?
• Does the logic hold?

Output Format

When using Action-Plan 150:

📋 **Action-Plan 150**

**Goal:** [Clear statement of what we're achieving]

**Steps:**
1. [ ] **Step 1:** [Action] → Success: [Criteria]
2. [ ] **Step 2:** [Action] → Success: [Criteria]
3. [ ] **Step 3:** [Action] → Success: [Criteria]
...

**Success Criteria:**
- ✅ Done when: [measurable outcome]
- 📊 Quality bar: [standards to meet]
- 🧪 Validation: [how to confirm]

**Risks & Mitigation:**
| Risk | Impact | Mitigation |
|------|--------|------------|
| [Risk 1] | High/Med/Low | [How to handle] |
| [Risk 2] | High/Med/Low | [How to handle] |

**Fallback Points:**
- 🔄 After Step X: Can safely stop here
- ↩️ Rollback plan: [How to undo if needed]

**Estimates:**
- Time: [Realistic estimate]
- Dependencies: [What's needed first]

Ready to execute?

Plan Quality Levels

Level	Characteristics	When to Use
Quick Plan	Goal + 3-5 steps	Simple tasks
Standard Plan	Steps + criteria + main risks	Normal work
Full Plan	Everything + estimates + alternatives	Complex/critical work

Operational Rules

1. NO BLIND EXECUTION: Don't start complex work without a plan
2. STEPS MUST BE ACTIONABLE: Each step = clear action
3. SUCCESS MUST BE MEASURABLE: Know when each step is done
4. RISKS MUST BE REAL: Think about what actually could fail
5. FALLBACKS MUST EXIST: Always have a way to stop/undo safely
6. PLANS CAN EVOLVE: Update as you learn more

Examples

❌ Without Action-Plan

Task: "Add user authentication"
Approach: "Just start coding"
Result: 3 days in, realized need for database changes, 
session handling unclear, no tests, rollback impossible

✅ With Action-Plan 150

Task: "Add user authentication"

📋 Action-Plan 150:

Goal: Users can register, login, logout securely

Steps:
1. [ ] Database: Add users table → Success: Migration runs
2. [ ] Backend: Auth endpoints → Success: API tests pass
3. [ ] Frontend: Login/register forms → Success: E2E tests pass
4. [ ] Security: Rate limiting, HTTPS → Success: Security audit pass
5. [ ] Deploy: Staged rollout → Success: No errors in 24h

Success Criteria:
- ✅ Users can register with email/password
- ✅ Login returns JWT token
- ✅ Protected routes reject unauthenticated
- 📊 Response time < 200ms
- 🧪 All auth tests green

Risks & Mitigation:
| Risk | Impact | Mitigation |
|------|--------|------------|
| Password leak | Critical | Use bcrypt, never log passwords |
| Session hijack | High | Secure cookies, short expiry |
| Brute force | Medium | Rate limiting, lockout |

Fallback Points:
- 🔄 After Step 2: Can deploy backend without frontend
- ↩️ Rollback: Feature flag to disable auth completely

Time Estimate: 2-3 days
Dependencies: Database access, frontend framework

Ready to execute!

✅ Quick Plan Example

Task: "Fix the button color"

📋 Quick Plan:
1. Find button component
2. Change color value
3. Test in browser
4. Commit

Success: Button is correct color, no side effects
Risk: Might affect other buttons → Check shared styles
Time: 15 minutes

Failure Modes & Recovery

Failure	Detection	Recovery
No plan	Coding without knowing goal	Stop, create plan first
Vague steps	"Make it work" type steps	Break into specific actions
No success criteria	"I think it's done"	Define measurable outcomes
Ignored risks	Surprised by problems	Add risk analysis
No fallback	Can't undo, can't stop	Define safe points

Relationship to Other Skills

Skill	Focus
goal-clarity-150	WHAT we want (the destination)
research-deep-150	WHAT we know (the information)
impact-map-150	WHAT it affects (the scope)
action-plan-150	HOW to get there (the path)
deep-think-150	HOW to reason about it
max-quality-150	HOW to do it well

Natural Flow:

goal-clarity → research-deep → impact-map → action-plan → execute
     ↓              ↓              ↓            ↓
  What we       What we        What's      How to
   want          know         affected     do it

Session Log Entry (MANDATORY)

After completing this skill, write to

.sessions/SESSION_[date]-[name].md

:

### [HH:MM] Plan-Action 150 Complete
**Action:** Created/Updated action plan
**Result:** <plan readiness>
**Artifacts:** <task.md/implementation_plan.md>

---

Remember: A plan isn't bureaucracy — it's a map. You can deviate from a map, but you can't deviate from nothing. Plans prevent the chaos of "figuring it out as we go."