Menu
Extract reusable patterns from the current session and store as Tekio adaptations or memories
The review-first prompt is selected by default. You can switch to a direct command or download a local copy.
Read SKILL.md and any companion files shown by SkillsMP before deciding whether to install.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
A direct command skips the review prompt. Inspect the source before running it.
npx skills add https://github.com/InugamiDev/ultrathink-oss --skill learn-patternThe command stays on one line. Scroll horizontally to inspect it before copying.
Prefer a local copy? Download the files currently available to SkillsMP.
Unified design foundations — design system architecture, tokens, component specs, visual principles, creative vision, figma integration, plus brand design system loader (66 real brands via DESIGN.md). Absorbs design, design-system, design-systems, design-principles, design-router, creative-vision, figma, design-md.
Render, summarize, and present markdown documents and structured content in multiple output modes
Ultra UI skill - combines Google's DESIGN.md spec (machine-readable design tokens) with the ui-ux-pro-max knowledge base (91 styles, 161 palettes, 73 font pairings, 161 products, 104 UX guidelines, 25 chart types). Generates lint-clean DESIGN.md files, validates token references and WCAG contrast, exports Tailwind/DTCG tokens, and diffs design systems version-over-version.
Based on SOC occupation classification
Showing SKILL.md
| name | learn-pattern |
| description | Extract reusable patterns from the current session and store as Tekio adaptations or memories |
| layer | utility |
| category | learning |
| triggers | ["learn from this","extract pattern","remember this pattern","save what we learned","learn","what did we learn"] |
| inputs | [{"context":"Current session context or specific interaction to learn from"},{"scope":"Project scope for pattern storage"}] |
| outputs | [{"patterns":"Extracted patterns with confidence scores"},{"storage":"Where patterns were saved (Tekio adaptation or memory)"}] |
| linksTo | ["debug","fix","refactor","sequential-thinking"] |
| linkedFrom | ["cook","audit","team"] |
| preferredNextSkills | ["verify","quality-gate"] |
| fallbackSkills | ["sequential-thinking"] |
| riskLevel | low |
| memoryReadPolicy | full |
| memoryWritePolicy | full |
| sideEffects | ["Creates Tekio adaptations in database","Creates memory entries in database"] |
Extract reusable engineering patterns from the current session and persist them for future sessions. Unlike Tekio wheel-turns (which learn from failures), this skill proactively captures successes, techniques, and insights mid-session.
Use this when:
| Type | Description | Storage | Example |
|---|---|---|---|
| Error Resolution | How a specific error was fixed | Tekio (defensive) | "TS2322 in Neon queries → cast with as Record<string, unknown>[]" |
| Debugging Technique | Systematic approach that worked | Memory (solution) | "Galaxy canvas memory leak → useMemo for filtered data + useRef for animation" |
| Project Convention | Discovered project patterns | Memory (pattern) | "All API routes use getDb() singleton, never inline neon import" |
| Architectural Decision | Design choices with rationale | Memory (decision) | "Chose pgvector + pg_trgm hybrid over pure vector search for memory recall" |
| Workaround | Known limitation with mitigation | Tekio (auxiliary) | "Bash set -u + empty arrays → use ${arr[@]+\"${arr[@]}\"} safe expansion" |
| Performance Insight | Optimization that worked | Memory (insight) | "Promise.all for independent DB queries cut response time 60%" |
Each extracted pattern gets a confidence score:
| Score | Meaning | Criteria |
|---|---|---|
| 0.9-1.0 | Proven | Verified by tests, applied 3+ times |
| 0.7-0.8 |
| High |
| Worked in this session, consistent with docs |
| 0.5-0.6 | Medium | Worked once, untested edge cases |
| 0.3-0.4 | Low | Hypothesis, not fully validated |
Is it about preventing a failure? → Tekio adaptation (defensive)
Is it about detecting issues early? → Tekio adaptation (auxiliary)
Is it about a better approach? → Tekio adaptation (offensive)
Is it a project-specific convention? → Memory (pattern/architecture)
Is it a reusable debugging technique? → Memory (solution)
Is it a design decision? → Memory (decision)
Examine recent work in the session:
For each pattern found, capture:
{
content: "Clear, actionable description of the pattern",
category: "solution" | "pattern" | "decision" | "architecture" | "insight",
importance: 1-10, // How broadly applicable
confidence: 0-1, // How well validated
scope: "project/name", // Where it applies
tags: ["#auto", "#learned", "#category"]
}
Before saving, check against existing knowledge:
# For memory entries
npx tsx memory/scripts/memory-runner.ts save '<json>'
# For Tekio adaptations (from corrections/failures)
npx tsx memory/scripts/memory-runner.ts wheel-correct '<wrong>' '<right>' [scope]
PATTERNS EXTRACTED: 3
1. [solution] Promise.all for parallel DB queries (confidence: 0.9, importance: 7)
→ Saved to memory: abc123
2. [defensive] Neon getDb() singleton prevents connection leaks (confidence: 0.8, importance: 8)
→ Saved as Tekio adaptation
3. [pattern] Dashboard API routes follow getDb() + try/catch + NextResponse pattern
→ Already exists (updated confidence 0.7 → 0.85)
| Pitfall | Impact | Fix |
|---|---|---|
| Saving trivial patterns | Memory pollution, low signal-to-noise | Filter: importance >= 5 for patterns |
| Missing the WHY | Pattern is remembered but not understood | Always include rationale and context |
| Over-confident scoring | False patterns get applied broadly | Start at 0.5, let repeated use increase confidence |
| Not deduplicating | Same pattern saved 5 times | Always search before saving |
| Too broad scope | Project-specific pattern applied globally | Scope patterns to project unless truly universal |
| Saving during exploration | Half-baked insights pollute memory | Only save after validation/verification |
Extracted pattern:
Content: "React canvas animations with filter state: use useMemo for filtered
data and useRef for values needed in animation loop. Never depend on state
directly in requestAnimationFrame — use refs to avoid teardown/rebuild."
Category: solution
Importance: 7
Confidence: 0.85
Tags: #react #animation #performance #memory-leak
Extracted pattern:
Content: "UltraThink dashboard API routes pattern: import getDb from @/lib/db,
wrap handler in try/catch, return NextResponse.json with proper status codes.
Never use inline neon() imports — they create connection pool issues."
Category: pattern
Importance: 8
Confidence: 0.9
Scope: ai-agents/ultrathink
Tags: #convention #api #database
Extracted Tekio adaptation:
Trigger: "SQL query with user input"
Rule: "Always use websearch_to_tsquery() instead of to_tsquery() for user-provided
search terms. to_tsquery() throws on special characters — websearch_to_tsquery()
handles them gracefully."
Category: defensive
Confidence: 0.95
Learn-pattern can be invoked: