Skip to main content
skill CX Codex

mastra-factory — Claude Code skill: operate Mastra Factory

A skill that drives Mastra Factory through its CLI: status, queues, projects and sessions, with mutations only inside your granted scope.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add mastra-ai/skills --skill mastra-factory -a claude-code
Open on GitHub

What mastra-factory does

This skill turns mastra api factory into the control plane for Factory. The agent inspects project state, work items, metrics, thresholds, decisions and supervisor health, then reports what is active, queued, blocked, which sessions are running and which decision needs a human.

Read-only is the default. Before any mutation the skill establishes your operating scope, which may cover a single action or grant standing autonomy over named projects and action types. Inside a clear scope the agent acts without asking before every change. It fetches current state first, makes the smallest in-scope change, then refetches and reports IDs, revisions and the final state.

Start with references/connection.md: check that the CLI is available and pin down your actual Factory URL. Platform-hosted deployments use mastra auth whoami, and if you are logged out the skill offers mastra auth login instead of silently opening a browser. Local and self-hosted authentication can differ. Stage changes go through transitions rather than metadata updates, and each one needs the current revision plus a fresh UUID request ID. The skill never reads .env, bearer tokens or saved login contents, never invents IDs or sessions, and never reaches for private HTTP routes to work around the CLI.

What it helps with

  • see what is running and what is queued
  • find blocked or unhealthy work items
  • review thread history and current agent memory
  • pull metrics, thresholds and decisions into one report
  • run a Factory operation within granted authority

Example prompts for the agent

  • Connect to my Mastra Factory and show the project status
  • Which work items are blocked and which decisions are waiting on me?
  • Show the thread history and memory for this work item

How to install mastra-factory

1Skills CLI (Vercel)

npx skills add mastra-ai/skills --skill mastra-factory -a claude-code

2Manually: put the files in place

~/.claude/skills/mastra-factory/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/mastra-factory/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about mastra-factory

How do I install mastra-factory in Claude?

Run npx skills add mastra-ai/skills --skill mastra-factory -a claude-code. The skill lands in ~/.claude/skills/mastra-factory/SKILL.md.

What does mastra-factory do?

It gives the agent a way to operate mastra api factory: status, queues, projects, work items, metrics, decisions and supervisor sessions. Everything stays read-only unless you grant a scope.

Do I need a deployed repository to use it?

No. An explicit mastra api --url "$FACTORY_URL" factory project list works from an empty directory, and no .mastra-project.json is required. Never assume a shared host, organization or project name.

Source text

SKILL.md · Apache-2.0
---
name: mastra-factory
description: "Operate and supervise Mastra Factory through `mastra api factory`. Use for first-time Factory connection and login, status or queue summaries, project and work-item inspection, thread history and memory, metrics, health, decisions, attention, supervisor sessions, and user-authorized autonomous or interactive Factory operations on hosted, local, remote, or self-hosted servers."
license: Apache-2.0
metadata:
  author: Mastra
  version: "1.0.0"
  repository: https://github.com/mastra-ai/skills
---

# Mastra Factory Supervisor

Use `mastra api factory` as the operational control plane for Factory.

## First use: connect before inspecting

Read [`references/connection.md`](references/connection.md). Check CLI availability, establish the user's actual Factory URL, and for platform-hosted deployments run `mastra auth whoami`. If logged out, offer `mastra auth login` rather than silently starting browser login. Local/self-hosted authentication may differ.

An explicit `mastra api --url "$FACTORY_URL" factory project list` works from an empty directory; no deployed repository or `.mastra-project.json` is required. Never assume a shared host, organization, project name, or ID. Preserve the explicit target on subsequent commands.

## Default behavior

For status, inspection, diagnosis, queue review, or recommendation requests:

1. Stay read-only.
2. Select the sole or explicitly named project; report choices when ambiguous.
3. Inspect project state, work items, metrics, thresholds, decisions, attention, and supervisor health/session.
4. Correlate stages, revisions, sessions, decisions, and health findings.
5. Report active/queued work, blocked or unhealthy items, running sessions, pending decisions, human attention, and one recommended next action.
6. Execute a recommendation only when the current request or a previously granted operating scope authorizes it.

## Required reference

Read [`references/factory-supervisor.md`](references/factory-supervisor.md) before running Factory commands for output control, contracts, the read-only workflow, mutation protocol, governance constraints, durable-session limitations, and error handling.

For “my work,” actual execution progress, thread messages, memory, or health interpretation, also read [`references/session-inspection.md`](references/session-inspection.md). Distinguish card stages from running agents, historical messages from current memory, and agent claims from verified repository outcomes.

## Safety boundary

- Never read or reveal `.env`, bearer tokens, saved login contents, or platform/provider credentials.
- Never invent IDs, stages, revisions, request IDs, or sessions.
- Establish the user's operating scope before mutating. Authorization may cover one action or grant standing autonomy over named projects, resources, action types, or objectives.
- Within a clear delegated scope, act without asking for confirmation before every mutation. Ask only when an action is ambiguous, outside scope, or materially more destructive than the granted authority.
- Fetch current state before a write, make the smallest in-scope change, then refetch and report IDs, revisions, and final state.
- Use transitions—not metadata updates—for stage changes, with the current revision and a fresh UUID request ID.
- Never use private HTTP routes to bypass unsupported CLI operations.
- If `work-item start` lacks a supported durable user session, report the block; never substitute the supervisor session or invent a session UUID.

© Mastra · Apache-2.0 · mastra-ai/skills

Similar skills

All skills for Agent setup: memory, context, orchestration →

From the same set2

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor