MinIO Agent Skills

Official Agent Skills published by MinIO, for any skills-compatible AI agent (Claude Code, Codex CLI, Cursor, Gemini CLI, GitHub Copilot, and more).

A skill is a folder with a SKILL.md (name + description + instructions) and, optionally, a references/ directory the agent loads on demand. Agents discover skills through progressive disclosure: the name/description are cheap to keep resident, and the full instructions load only when a task matches.

This repository is the public home for all MinIO skills — install any of them with one command, no MinIO source checkout required.

Available skills

Skill	Install	What it does
aimem	npx skills add minio/skills/aimem	Work inside an AIMem workspace — locate the mount, durability semantics, agent memory, object metadata/annotations, and secrets.
mc	npx skills add minio/skills/mc	Drive mc, the AIStor command-line client — connect via an alias, list/copy/remove objects and buckets, and run bucket config, replication, lifecycle, cluster admin, and Tables operations non-interactively with JSON-lines output.

Installing

These skills follow the open Agent Skills standard, so you can install them with the skills CLI:

# add a single skill to your agent's skill directory
npx skills add minio/skills/aimem

# discover what's here
npx skills find minio

Or install manually — copy a skill folder into your agent's skill directory:

# Claude Code
cp -r aimem ~/.claude/skills/aimem
# Codex CLI, Cursor, Gemini CLI, Copilot
cp -r aimem ~/.agents/skills/aimem

When do I need to install manually?

AIMem installs the aimem skill into your home skill directories automatically on every mount, so if your agent's reasoning runs inside the sandbox where AIMem is mounted, you do not need this repo.

Install from here when your agent's reasoning runs outside the sandbox — for example the OpenAI Agents SDK or other managed-agent harnesses, where the model lives in a control plane and an in-sandbox install is invisible to it. Add the skill host-side, alongside your own skills, before the run starts.

Contributing a skill

Each top-level directory is one skill and must contain a spec-compliant SKILL.md:

• name — lowercase, digits, hyphens; must match the directory name.
• description — what it does and when to use it (≤ 1024 chars).

See the Agent Skills specification. Validate locally with skills-ref validate ./<skill>.

License

Apache-2.0 — see LICENSE.