All Skill Categories

Image, video, audio and design skills — generation prompts, editing pipelines, transcription and asset preparation. Most pair a creative brief with a concrete production step.

Code-facing skills: review checklists, refactoring passes, test generation, migration scripts and language-specific conventions. Most encode a team's standards so an agent applies them the same way every time.

Memory and knowledge skills: what an agent should remember, how to store it, and how to retrieve it later without dragging stale context into every conversation.

Skills for reading and producing documents — PDFs, Word files, slide decks, spreadsheets and long-form reports. They cover both directions: pulling structure out of a file, and writing a finished one back.

Analysis skills that turn raw tables into answers — SQL generation, dataset profiling, chart specifications and reporting pipelines. Useful when the agent needs a method, not just database access.

Skills that teach an agent how a specific service expects to be used — Slack, Notion, Jira, GitHub, Linear and other APIs. They add conventions and etiquette on top of raw tool access.

Security skills for authorized work — code auditing, dependency review, secret scanning and compliance checklists. They encode a review method, not an attack.

Browser skills: navigating pages, filling forms, extracting structured data and driving automation frameworks. They matter when the work only exists behind a UI.

Skills that change how an agent plans, delegates and reports — sub-agent orchestration, review loops, planning templates and persona definitions. Reach for these when the problem is the agent's process rather than a missing tool.

Everyday work skills: inbox triage, meeting notes, task breakdowns, writing assistance and personal workflows. These are the ones people install first and keep enabled.

Skills that do not fit the main categories yet. Worth browsing when you know what you need but not what it would be filed under.

Research skills that decide where to look, how deep to go and when to stop — multi-source sweeps, source evaluation and structured summarization.