Oracle Incubate Skill

Track knowledge as it grows from observation to wisdom

Purpose

Oracle-incubate monitors knowledge maturation and suggests when learnings are ready to "level up" in the maturity hierarchy.

Knowledge Maturity Levels

code

🥒 Observation  →  🌱 Learning  →  🌿 Pattern  →  🌳 Principle  →  🔮 Wisdom
     (raw)          (tested)       (repeated)      (universal)      (core)

Proactive Triggers

MUST Use Incubate When:

Session End:

•After significant work session
•After retrospective (rrr)
•User says "wrap up", "end session"

Pattern Recognition:

•User says: "this keeps happening", "I always do this"
•User says: "we keep learning this", "another time"
•Same topic appears 3+ times in Oracle

Promotion Signals:

•User says: "this is definitely a pattern now"
•User says: "we should remember this"
•User says: "level up", "mature", "promote"

SHOULD Use Incubate When:

•Reviewing past learnings
•After successful problem resolution
•When consolidating session insights

Incubation Workflow

1. Scan for Maturation Candidates

javascript

// Search for learnings that might be ready to level up
oracle_search({
  type: "learning",
  limit: 20
})

Look for:

•Learnings referenced multiple times
•Learnings proven in different contexts
•Learnings that led to decisions

2. Suggest Promotions

Present findings to user:

markdown

## 🌱 Incubation Report

### Ready for Promotion?

| Learning | Current | Times Used | Suggest → |
|----------|---------|------------|-----------|
| "Always validate webhooks" | 🌱 | 5x | 🌿 Pattern |
| "Subagents for bulk work" | 🌿 | 10x | 🌳 Principle |

### Still Incubating

- "New API auth pattern" (🥒, tested 1x)
- "TypeScript strict mode" (🌱, tested 2x)

3. Promote with Confidence

javascript

oracle_learn({
  pattern: "Existing pattern with refinement",
  stage: "pattern",        // promoted from "learning"
  confidence: "high",
  times_validated: 5,
  teachable: true,
  promotion_date: "2026-01-02",
  promoted_from: "learning"
})

Maturity Criteria

From	To	Criteria
🥒 → 🌱	Observation → Learning	Worked once, not disproven
🌱 → 🌿	Learning → Pattern	Used 3+ times, consistent results
🌿 → 🌳	Pattern → Principle	Context-independent, universally true
🌳 → 🔮	Principle → Wisdom	Changed behavior fundamentally

Dashboard Command

When user asks "show knowledge maturity" or "incubation status":

markdown

## 🌡️ Knowledge Maturity Dashboard

### By Stage
| Stage | Count | Ready to Promote |
|-------|-------|------------------|
| 🥒 Observations | 15 | 3 (tested once) |
| 🌱 Learnings | 42 | 8 (3+ uses) |
| 🌿 Patterns | 23 | 2 (universal) |
| 🌳 Principles | 12 | 1 (core) |
| 🔮 Wisdom | 5 | - |

### Recent Activity
- 🌱→🌿 "Subagent delegation" promoted (Dec 2025)
- 🥒→🌱 "TypeScript strict mode" validated (Jan 2026)

### Oldest Unreviewed
- "API versioning strategy" (🌱, 45 days)
- "Error boundary pattern" (🌱, 30 days)

Integration with Oracle

Tool	Incubate Role
`oracle_learn`	Add maturity metadata when capturing
`oracle_search`	Find promotion candidates
`oracle_consult`	Consider maturity level in guidance

Incubation Prompts

At Session End

code

"Before we wrap up, let me check Oracle for maturing knowledge..."

[Run incubation scan]

"Found 2 learnings that might be ready to promote. Would you like to review?"

After Multiple Uses

code

"I noticed we've used the 'subagent for bulk work' pattern 5 times now.
Currently marked as 🌱 Learning. Promote to 🌿 Pattern?"

On Manual Request

code

"Showing knowledge maturity dashboard..."
[Display dashboard]
"3 items ready for promotion review."

Future: Automation Hooks

Phase B establishes tracking. Future phases may add:

•Auto-promotion thresholds (3 uses = auto-suggest)
•Maturity decay (unused patterns demote)
•Cross-project pattern detection
•Teaching material generation (Phase C)

Quick Reference

User Says	Action
"incubation status"	Show dashboard
"what's maturing?"	List promotion candidates
"promote this to pattern"	Update with oracle_learn
"wrap up session"	Run incubation scan
"this keeps happening"	Check if pattern exists, suggest capture