Skills Library

aibox ships 84 curated skills across 14 categories. Every skill follows the open SKILL.md standard and is deployed into .claude/skills/ based on your process packages and addons.

What Are Skills?

A skill is a directory containing a SKILL.md file (and optional references/ files) that teaches an AI agent how to perform a specific task. Skills use progressive disclosure:

1. Metadata (~100 tokens) -- name and description loaded at startup for all skills
2. Instructions (<5000 tokens) -- full SKILL.md body loaded when the skill activates
3. References (on demand) -- detailed reference files loaded only when needed

.claude/skills/
├── kubernetes-basics/
│   ├── SKILL.md              # Main instructions
│   └── references/
│       ├── cluster-architecture.md
│       ├── resource-cheatsheet.md
│       └── troubleshooting.md
├── code-review/
│   └── SKILL.md              # Simple skill, no references
└── ...

Categories at a Glance

Category	Skills	Description
Process	9	Backlog, decisions, standups, releases, incidents, retrospectives, agent coordination
Development	11	Code review, testing, debugging, refactoring, error handling, documentation
Language	7	Python, Rust, TypeScript, Go, Java, SQL style, LaTeX
Infrastructure	10	Docker, Kubernetes, DNS/networking, Terraform, Linux, shell scripting, CI/CD
Architecture	4	Software architecture, DDD, event-driven, system design
Design & Visual	7	Frontend, Tailwind, Excalidraw, infographics, logos, PixiJS, mobile UX
Data & Analytics	5	Data science, pipelines, visualization, feature engineering, data quality
AI & ML	6	AI fundamentals, RAG, prompt engineering, LLM evaluation, embeddings, ML pipelines
API & Integration	4	REST API design, GraphQL, gRPC/Protobuf, webhooks
Security	5	Auth patterns, secure coding, threat modeling, dependency audit, secrets
Observability	4	Logging, metrics, distributed tracing, alerting
Database	4	SQL patterns, data modeling, NoSQL, migrations
Performance	4	Profiling, caching, concurrency, load testing
Framework & SEO	5	FastAPI, Reflex, pandas/polars, Flutter, SEO

How Skills Are Deployed

Skills are deployed based on three sources, merged in order:

1. Process Packages

Your [process].packages config determines the base skill set. Each package bundles related skills:

[process]
packages = ["managed"]  # Expands to: core, tracking, standups, handover

The 13 packages and their skills:

Package	Skills
core	agent-management, owner-profile
tracking	backlog-context, decisions-adr, event-log, context-archiving
standups	standup-context
handover	session-handover, inter-agent-handover
product	estimation-planning, retrospective
code	code-review, testing-strategy, debugging, refactoring, tdd-workflow, error-handling, git-workflow, integration-testing
research	data-science, data-visualization, feature-engineering
documentation	documentation, latex-authoring
design	excalidraw, infographics, logo-design, frontend-design, mobile-app-design
architecture	software-architecture, system-design, domain-driven-design, event-driven-architecture
security	secure-coding, threat-modeling, dependency-audit, auth-patterns, secret-management, dependency-management
data	data-pipeline, data-quality, pandas-polars, embedding-vectordb
operations	ci-cd-setup, dockerfile-review, container-orchestration, logging-strategy, metrics-monitoring, incident-response, alerting-oncall, performance-profiling

4 convenience presets expand to multiple packages:

Preset	Expands to
managed	core, tracking, standups, handover
software	core, tracking, standups, handover, code, architecture
research-project	core, tracking, standups, handover, research, documentation
full-product	core, tracking, standups, handover, code, architecture, design, product, security, operations

2. Addon Skills (automatic)

When you add an addon, its recommended skills are automatically included. No manual [skills].include needed:

Addon	Auto-deployed skills
python	python-best-practices, fastapi-patterns, pandas-polars
rust	rust-conventions, concurrency-patterns
go	go-conventions, concurrency-patterns
node	typescript-patterns, tailwind
latex	latex-authoring, documentation
typst	documentation
kubernetes	kubernetes-basics, container-orchestration
infrastructure	terraform-basics
docs-* (all 6)	documentation

For example, adding the python addon to a managed project automatically deploys python-best-practices, fastapi-patterns, and pandas-polars alongside the managed package skills.

3. Manual Include/Exclude

Fine-tune with [skills] in aibox.toml:

[skills]
include = ["flutter-development", "seo-optimization"]  # Add extra skills
exclude = ["standup-context"]                            # Remove unwanted ones

Core skills (agent-management, owner-profile) cannot be excluded.

Managing Skills via CLI

# List all skills and their deploy status
aibox skill list

# Add a skill (updates aibox.toml skills.include)
aibox skill add flutter-development

# Remove a skill (adds to skills.exclude)
aibox skill remove standup-context

# View skill details
aibox skill info rust-conventions

Updating Skills

When you upgrade aibox, new skills are added automatically on the next aibox sync. Existing skills are never overwritten -- only missing ones are created.

Custom Skills

Create your own skills alongside the bundled ones:

# .claude/skills/my-custom-skill/SKILL.md
---
name: my-custom-skill
description: What this does and when to use it. Be specific to help trigger detection.
allowed-tools: Bash(npm:*) Read Write
---

# My Custom Skill

## When to Use
Describe trigger conditions.

## Instructions
Step-by-step agent instructions.

## Examples
Scenario-based examples.

SKILL.md Format

Every skill follows the Agent Skills specification:

Field	Required	Description
name	Yes	Lowercase kebab-case, max 64 chars
description	Yes	What it does AND when to use it, max 1024 chars
allowed-tools	No	Pre-approved tools (e.g., Bash(kubectl:*) Read Write)
license	No	License identifier
compatibility	No	Environment requirements
metadata	No	Arbitrary key-value pairs

Progressive Disclosure Best Practices

• Keep SKILL.md under 500 lines (ideally 50-150)
• Move detailed reference material to references/*.md
• Agents load references on demand, keeping context lean
• Include 2-4 scenario-based examples in every skill

Security

Only install skills from trusted sources

Skills contain instructions that AI agents execute. A malicious skill could instruct an agent to modify files or exfiltrate data. Only install skills from sources you trust.

• Review before installing -- read the SKILL.md file before adding third-party skills
• Version control -- commit .claude/skills/ to git so changes are tracked
• Prefer bundled -- the 84 skills shipped with aibox are curated and maintained
• Audit external skills -- treat them like any dependency