Activate for The Side System™ AI strategy, data analysis, design systems, and product design work. This is the technical backbone — feeds into most TSS solutions. Use for: AI integration recommendations, data pattern analysis, design system architecture, product design decisions, and technology stack guidance within any TSS engagement.
From brand-bond-osnpx claudepluginhub brandbondco/brand-bond-os --plugin brand-bond-osThis skill uses the workspace's default tool permissions.
Guides Next.js Cache Components and Partial Prerendering (PPR) with cacheComponents enabled. Implements 'use cache', cacheLife(), cacheTag(), revalidateTag(), static/dynamic optimization, and cache debugging.
Migrates code, prompts, and API calls from Claude Sonnet 4.0/4.5 or Opus 4.1 to Opus 4.5, updating model strings on Anthropic, AWS, GCP, Azure platforms.
Details PluginEval's skill quality evaluation: 3 layers (static, LLM judge), 10 dimensions, rubrics, formulas, anti-patterns, badges. Use to interpret scores, improve triggering, calibrate thresholds.
AI on a broken system builds a faster broken system.
This skill exists to prevent that. Every technical recommendation is grounded in the actual problem, the available data, and the realistic maintenance capacity of the client.
| Trigger | Domain |
|---|---|
| Client has operational data that hasn't been analyzed | Data |
| Pattern detection needed during side:sight | Data + AI |
| AI tool recommendation requested for an engagement | AI Strategy |
| Design system needs to be built or audited | Design Systems |
| Product UX/flow decision required | Product Design |
| Client's tech stack needs assessment | AI + Product |
| Any claim involving "AI will solve this" | AI Strategy — validate first |
Before recommending any AI solution:
| Question | Required answer |
|---|---|
| What specific task is manual, slow, or error-prone? | Named, not vague |
| Is there enough clean data? | Volume + quality assessment |
| Can this integrate with what already exists? | Stack compatibility |
| Who owns this after we disengage? | Maintenance plan |
| What breaks if the AI fails or produces wrong output? | Failure mode analysis |
Hard rule: If any of these has no answer, do not recommend the AI solution yet. Define what's missing first.
| Phase | AI application | Tool options |
|---|---|---|
side:sight | Pattern recognition, behavioral signals, anomaly detection | Claude, Python/pandas, Metabase |
side:scope | Priority scoring, scenario modeling, constraint mapping | Claude, spreadsheet models |
side:setup | Workflow automation, tool integration, process templating | Make, n8n, Zapier |
side:scale | Growth simulation, AI-augmented operations, capacity modeling | Claude, custom LLM workflow |
side:sync | Continuous monitoring, drift detection, health alerts | Mixpanel, PostHog, custom scripts |
| Category | Options | When to recommend |
|---|---|---|
| Language / reasoning | Claude (primary), GPT-4o | Content, analysis, structured output |
| Automation | Make, n8n, Zapier | Workflow connections, repetitive processes |
| Data analysis | Python + pandas, Metabase, Tableau | Pattern detection, reporting layer |
| AI ops | LangChain, LlamaIndex | Custom RAG systems, document intelligence |
| Design AI | Figma AI, Midjourney | Concept exploration, not final production |
| CRM intelligence | Clay, Apollo, HubSpot AI | Lead enrichment, sales system layer |
| Product analytics | Mixpanel, Amplitude, PostHog | User behavior, funnel analysis |
Every data finding must be structured as:
Pattern: [Observable fact — what the data shows]
Signal: [What this pattern means]
Implication: [What breaks if this continues]
Recommendation: [First specific action]
| Source | Pattern to look for |
|---|---|
| Revenue timeline | Stall point, growth ceiling, anomaly, seasonality |
| Customer retention | Churn trigger, loyalty driver, segment behavior |
| Team output data | Bottleneck, productivity gap, capacity ceiling |
| Sales pipeline | Conversion drop-off, cycle length, win/loss signal |
| Product/web analytics | Friction point, drop-off, behavioral pattern |
| Financial ratios | Cash efficiency, burn rate, margin trend |
If the client doesn't have clean data, that's itself a side:sight finding:
Pattern: Company has no structured performance data
Signal: Decisions are being made on intuition or outdated reports
Implication: Cannot validate whether growth initiatives are working
Recommendation: side:setup priority — implement a lightweight data layer first
| Layer | What it contains |
|---|---|
| Tokens | Color, spacing, typography, radius, shadow — all as variables |
| Primitives | Base components: Button, Input, Badge, Icon |
| Patterns | Composed components: Form, Card, Modal, Navigation |
| Templates | Page-level compositions |
{category}-{role}-{variant}
Examples:
color-surface-primary
color-text-secondary
color-border-default
spacing-component-gap
spacing-layout-section
typography-body-size
typography-heading-weight
radius-component-default
Before any component or system is accepted:
When auditing an existing design system:
1. TOKEN AUDIT — Are values hardcoded or tokenized?
2. COMPONENT AUDIT — Are states complete? Are variants logical?
3. CONSISTENCY — Are the same patterns used across the system?
4. DOCUMENTATION — Can a new designer/developer use this without asking?
5. MAINTENANCE — Is there a process for updating the system?
| Question | Standard |
|---|---|
| What is the user's single goal on this screen? | One primary action |
| Cognitive load? | Fewer than 5 decision points visible |
| Platform pattern? | iOS: HIG / Android: Material 3 — never fight platform |
| Performance? | All animation at 60fps before shipping |
| Accessibility? | 44×44pt min hit target, AA contrast, screen reader tested |
| Pattern | Key principles |
|---|---|
| Dashboard | Data density without overload — hierarchy, not decoration |
| Multi-step flows | Progress indication, state persistence across steps |
| Role-based access | Visible permission states, clear locked vs. unlocked UI |
| Empty states | Never blank — explain what goes here and how to start |
| Onboarding | Progressive disclosure, activation milestones, no feature dump |
Feature request received
↓
Check if pattern exists in design system
↓
Exists → Use it. No one-off components.
Doesn't exist → Create token-level primitive first
↓
Document in system before shipping
↓
No exceptions
| Skill | What tss-intel provides |
|---|---|
tss-sight | Data pattern analysis, behavioral signals |
tss-scope | Technology feasibility, AI effort estimation |
tss-proposal | Stack assessment, AI integration cost/complexity |
tss-deck | Technical accuracy review for AI/product claims |
tss-copy | Technical precision review for any AI or product claims |
tss-competitive | Tech stack monitoring for competitors |