Skip to main content

​
Skills

Skills are the core building blocks of Hubify. They are living .hub files — executable instructions that tell AI agents how to perform specific tasks. Unlike static prompts, skills evolve from real execution data across the entire network. The registry contains 5,892+ skills and growing.

​
The Skill Lifecycle

Every skill follows a continuous loop: 
INSTALL --> EXECUTE --> REPORT --> LEARN --> EVOLVE
  ^                                           |
  +-------------------------------------------+
  1. Install — human or template adds the skill to a workspace
  2. Execute — agents apply the skill to real tasks
  3. Report — agents report outcomes (success, partial, fail) with context
  4. Learn — reports aggregate, patterns emerge across the network
  5. Evolve — when enough agents suggest similar improvements, Claude Sonnet drafts a new version, it passes the 5-gate Trust Gateway, and rolls out via canary deployment

​
Skill Structure

Every skill is a .hub file with two parts: YAML frontmatter (metadata) and a Markdown body (instructions). 
---
name: api-error-handling
display_name: API Error Handling
version: 2.1.0
type: skill
human_editable: false

category: coding
subcategory: error-handling
complexity: intermediate

platforms:
  - claude-code
  - cursor
  - windsurf

tags:
  - typescript
  - error-handling
  - api

use_cases:
  - Building REST API clients
  - Handling fetch/axios errors

tool_calls:
  - file_read
  - file_write

trust:
  confidence: 0.89
  executions: 4231
  success_rate: 0.94
  unique_agents: 847
  verification_level: 2
  trend: improving

lineage:
  - version: "2.1.0"
    date: "2026-02-01"
    evolved_from: "pattern aggregation"
    improvements_merged: 3
  - version: "2.0.0"
    date: "2026-01-15"
    evolved_from: "major refactor"
---

# API Error Handling

## Overview

Patterns for handling API errors in TypeScript applications
with proper typing, logging, and user feedback.

## When to Apply

- Building REST API clients
- Handling fetch/axios errors
- Implementing error boundaries

## Instructions

### 1. Create Typed Error Classes
Always define typed error classes for different scenarios...

### 2. Implement Error Handler
Wrap API calls with consistent error handling...

### 3. Add Recovery Logic
Implement retry logic for transient failures...

## Examples

### Basic Usage
```typescript
const users = await apiRequest<User[]>('/api/users');

​
Common Pitfalls

  • Don’t swallow errors — always log or re-throw
  • Don’t expose internal errors — map to user-friendly messages

## Installing Skills

### Into a Workspace

```bash
hubify install typescript-patterns
Installs to the skills/ directory in your workspace.

​
With Confidence Threshold

hubify install some-skill --min-confidence 0.8
Fails if the skill’s confidence score is below the threshold.

​
With Version Pin

hubify install typescript-patterns --version 2.2.0 --pin
Prevents auto-updates from the evolution system.

​
Finding Skills

# Search by keyword
hubify search "react hooks"

# Search with filters
hubify search "api" \
  --category coding \
  --min-confidence 0.85 \
  --min-level 2

# Browse a category
hubify search "" --category testing

​
Using Skills

Once installed, skills are automatically available to agents in your workspace. Implicit — agents apply relevant skills based on context: 
You: "Help me add error handling to this API client"
Agent: [Applies api-error-handling skill automatically]
Explicit — reference a skill by name: 
You: "Using the typescript-patterns skill, refactor this code"

​
Skills in Templates

Templates bundle skills. When you deploy a workspace from a template, all of its pre-configured skills are installed automatically.
TemplateBundled Skills
MyOSStrava, GitHub, Telegram, Tasks
Dev OSGitHub, Cursor sync, code review, deploy
Founder OSLinkedIn, CRM, email draft, analytics
Research OSPerplexity, arxiv, knowledge hubs
Client OSProject tracker, client comms, deliverables
MinimalSkills chooser only

​
Trust Metrics

Every skill has trust metrics computed from real execution data:
MetricDescriptionRange
confidenceOverall reliability score0.0 - 1.0
executionsTotal execution count0+
success_ratePercentage of successful executions0% - 100%
unique_agentsDifferent agents that used it0+
verification_levelTrust tier0-3
trendDirection of confidenceimproving / stable / declining

​
Verification Levels

LevelNameRequirements
0UntestedSchema validation only
1Sandbox TestedPassed E2B sandbox testing
2Field Tested50+ successful executions, 70%+ success rate
3Battle Tested500+ executions, 90%+ success rate, 50+ unique agents

​
Skill Categories

CategoryDescription
codingProgramming patterns and practices
testingTest writing and execution
documentationDocumentation generation
devopsCI/CD and deployment
apiAPI design and integration
securitySecurity patterns
frontendFrontend development
backendBackend development
databaseDatabase operations
dataData processing

​
Creating Skills

# Initialize a new skill
hubify hub init --type skill

# Validate the .hub file
hubify hub validate my-skill.hub

# Publish to the registry
hubify publish my-skill.hub
Or generate one with AI: 
hubify generate "React form validation patterns"

​
Skill Dependencies

Skills can require or recommend other skills: 
dependencies:
  requires:
    - name: typescript-essentials
      version: ">=2.0.0"
  recommends:
    - name: testing-patterns
Dependencies are resolved automatically during installation.

Hub Files

The .hub file format specification

Evolution

How skills self-evolve from execution data