Spuree API Skills

Spuree provides an agent-friendly cloud storage via a simple Project → Folder → File hierarchy. Agents authenticate with an API key, then create projects, organize folders, and upload/download files — all through REST APIs designed for programmatic access.

These skills are structured API documentation for AI agent consumption. Each skill documents a specific API domain with endpoints, request/response schemas, and working examples.

Installation

npx skills add https://github.com/CheehooLabs/SpureeSkills

To update to the latest version, run the same command again.

Getting Started

After installing or connecting Spuree, ask your agent:

"How do I use Spuree skills?"

The getting-started skill explains the available public capabilities and offers a guided, read-only first step. You can also ask “What can Spuree do?” or “Walk me through Spuree.” The guide adapts to full skill/MCP clients and to read-only web connectors that expose search and fetch. Studio also publishes the same guide as the getting_started MCP tool for Claude Code on the web and ChatGPT.

Install Guide

Connect Claude Code, Claude Code Web & Cowork, ChatGPT, Codex, or OpenClaw — via the OAuth connector or skills + API keys. See INSTALL.md.

Available Skills

Skill	Description
Getting Started	Learn what Spuree can do and complete a safe guided first run
First Post	Take a new workspace from nothing to a reviewable post draft with your own agent over the MCP connector — a business, a product or a personal brand; your own material or a brief you approve, a quote before the first paid step, one confirmed brand memory build, a reel studied for pacing, a short vertical video, the draft under the brand
Authentication	Obtain JWT tokens, manage API keys
Project Management	Create, list, update, delete, and share projects; browse contents
Folder Management	Find canonical folders in one bounded call, list recent folders, and create, update, delete, browse, or download folder content
File Management	Get, create, upload, update, and delete files with checksum-verified uploads (includes name search)
File Comments	Add, list, resolve, update, and delete review comments on files — anchored to a line range, a video timecode, or a still — with drawings, threaded replies, and @mentions
Project Invitations	Manage sharing invitations for non-workspace members
Usage Report	Retrieve daily usage reports for a workspace or organization, broken down by API key, MCP client, and user, with allowance tracking
Workspace Management	List a workspace's members, add a member with a role, change a role, remove a member
Report Feedback	Send the Spuree team a bug report, feedback or a feature request as the person, after they approve the exact text, and check the status of their own reports — over the MCP connector

Authentication

All skills use the V1 API. Two authentication methods are supported:

Method	Header	Use case
JWT token	Authorization: Bearer <token>	Interactive sessions (1 hour expiry, refreshable)
API key	X-API-Key: <key>	Automation and long-lived access

See the Authentication skill for details on obtaining tokens and managing API keys.

Contributing

See CONTRIBUTING.md for how to edit a skill, the checks to run, and what a pull request needs before it can merge.

Contract checks and diagnostics

The folder-discovery contract is checked offline so documentation drift fails before release. The guard pins exact endpoint headings, ordered query parameters (types, required flags, defaults, enums, and bounds), canonical context object shapes, response field names and types, closed response enums, status codes, and bounded fallback behavior:

node scripts/check-folder-discovery-contract.mjs
node --test test/*.test.mjs

To diagnose duplicate public and plugin skill copies without changing or deleting any installation, run the script from a checkout of this repository and point --target at the project or agent workspace whose active catalog you want to inspect:

node scripts/diagnose-spuree-skill-copies.mjs --target /path/to/project
node scripts/diagnose-spuree-skill-copies.mjs --target /path/to/project --json

When the checkout and target project differ, the checkout remains the immutable source reference while --target controls project/workspace discovery. The diagnostic scans every public skill across .agents, .codex, .claude, OpenClaw workspace/global, Hermes global, and Codex plugin-cache roots. It reports paths, SHA-256 hashes, catalog scan order, exposed names, and plugin namespace/version metadata. Collision groups use the exposed skill name, so a namespaced plugin such as internal-spuree-skills:folder-management is not misreported as shadowing the public folder-management skill. If a cached plugin manifest is missing or unreadable, the diagnostic assigns that cache root a stable unknown-plugin-<hash> namespace instead of assuming the skill is bare. Scan order is informational; the active client owns runtime precedence.