Mastra AI Skills - Setup Guide¶

Source: mastra-ai/skills GitHub: mastra-ai/skills Skills: 5 skills · 32.6K total installs Category: AI Agent Framework First Seen: catalogued August 18, 2026 sweep (mastra on skills.sh since January 28, 2026) Quality Tier: 🟡 Trusted - official Mastra org; flagship mastra passes Gen Agent Trust Hub and Socket, Snyk Warn (named)

The official Mastra skill suite is the canonical reference for building AI applications with the Mastra TypeScript framework. Its core doctrine is version discipline: never trust training-data knowledge of Mastra APIs, always verify against the embedded docs in node_modules/@mastra/*/dist/docs/ or the remote https://mastra.ai/llms.txt before writing code. The flagship mastra skill carries 32.3K of the suite's 32.6K installs.

Installation¶

npx skills add mastra-ai/skills

Individual skills install with the explicit repo form:

npx skills add https://github.com/mastra-ai/skills --skill mastra

Prerequisites¶

Requirement	Details
Node.js + npx	For the skill installer
Mastra project	An existing TypeScript project with `@mastra/*` packages in `node_modules`
ES2022 modules	Required in TypeScript config for the framework

What It Provides¶

Skill	Installs	Purpose
mastra	32.3K	Reference guide for building agents and workflows with current Mastra APIs
create-mastra	90	Scaffolding new Mastra projects
mastra-best-practices	89	Framework conventions and patterns
mastra-embedded-docs-look-up	67	Lookup strategy against the installed version's embedded docs
mastra-embeded-docs-look-up	41	Earlier-named variant of the embedded docs lookup

The flagship covers the core building blocks: Agents (autonomous decision-making), Workflows (structured sequences), Tools (capability extension), Memory (context maintenance), and RAG (external knowledge). It also documents the model format "provider/model-name" (for example "openai/gpt-5.4"), setup guidance, common errors, and migration notes matched to the installed version.

Quick Start¶

Install: npx skills add mastra-ai/skills
Check the installed packages: ls node_modules/@mastra/
Ask the agent to build an agent or workflow - the skill routes every API question through the embedded docs of the exact installed version

CorpusIQ Use Cases¶

Use Case	How
Version-accurate agent builds	Kills the stale-training-data failure mode when generating Mastra agents and workflows
RAG and memory patterns	First-party reference for the Memory and RAG blocks we use in agent architecture
Workflow orchestration	Structured-sequence patterns translate to our multi-step automation design

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects

Limitations / Verification¶

Security audits on the mastra flagship: Gen Agent Trust Hub Pass, Socket Pass, Snyk Warn (named in the tier)
Publisher-page total verified (32.6K across 5 skills); 72 GitHub stars on the skills repo as of the sweep
create-mastra, mastra-best-practices, and the two embedded-docs variants are 41-90 installs each - early content
The two embedded-docs skills overlap (one is a typo-named earlier variant)
Additional source (Oct 10, 2026 sweep): the framework repository mastra-ai/mastra also ships agent skills under .claude/skills/ (30 indexed listings on skills.sh, ~12,987 combined installs; top: react-best-practices 2,810, ralph-plan 2,619, smoke-test 2,493). Distinct from the mastra-ai/skills suite above - check both when searching for Mastra skills.

npx skills add mastra-ai/skills   # verify install works

Powered by CorpusIQ