Eighty Mcp Examples — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Eighty Mcp Examples (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
A public Model Context Protocol server that lets any MCP-compatible client — Claude Desktop, Claude.ai Connectors, Cursor, ChatGPT Apps, MCP Inspector — read and write a real-life RPG character sheet.
The server is live at `https://mcp.80l.app`. This repo is the documentation, example prompts, and reference screenshots — not the server source.
MCP Inspector connected to mcp.80l.app, showing the eighty tool catalog
eighty is a gamified self-improvement iOS app: real-world quests, character stats, skill trees, badges, streaks, leaderboards, social challenges. The MCP server exposes the same 35 actions the iOS app uses — browse_quests, create_quest, take_quest, complete_quest, get_profile, send_challenge, view_public_profile, the whole catalog — so an LLM client can drive the experience: plan your day as quests, coach you through them, send challenges to friends, react to your social feed, level up your character.
It's the same backend the iOS app talks to. No fork. Your MCP session and your iOS app are two views over one account.
Add to your claude_desktop_config.json:
{
"mcpServers": {
"eighty": {
"url": "https://mcp.80l.app"
}
}
}Restart Claude Desktop. On first tool call, OAuth opens in your browser — sign in with Google, Apple, or email/password. (New users get an account auto-provisioned at sign-in.) You're now connected.
Full setup for other clients (Cursor, Claude.ai web Connectors, MCP Inspector, ChatGPT Apps) is in docs/auth.md.
Inspector — Anthropic's reference MCP client — runs locally with no setup:
npx @modelcontextprotocol/inspectorIn the browser: set Transport Type to Streamable HTTP, URL to https://mcp.80l.app, Connection Type to Via Proxy, click Connect. Authorize via the OAuth popup. The full tool catalog loads in the sidebar.
35 tools across 7 categories — quests, quest creation, skills, profile, social, challenges, leaderboard. Plus 4 ambient resources and 3 pre-built prompts. Full reference in docs/tools-reference.md.
A few signature examples:
| Tool | What it does |
|---|---|
view_public_profile | Same rich profile shape the iOS app shows on tap (level, stats, streaks, badges, completed quests, follower counts). |
create_quest + publish_quest | Author a real quest from MCP that ships into eighty's catalog after review. |
send_challenge (Premium) | Real-time challenge to another user with a time limit — fires a push notification to their phone. |
whats_next (prompt) | One-shot personalized recommendation drawn from your profile, active quests, favorite skills, and today's daily quest. |
Hand-picked example prompts (copy-paste into your client) are in prompts/.
client → mcp.80l.app (OAuth 2.1) → same backend handlers the iOS app uses → persistenceThe MCP server is a thin authorization shell — every tool delegates to the same handler the iOS app calls. One source of truth for business rules (quest caps, premium gating, social-graph privacy, idempotency). See docs/architecture.md.
OAuth 2.1 + PKCE + Dynamic Client Registration (RFC 7591). No API keys, no developer-portal signup. Any spec-compliant MCP client auto-discovers the OAuth endpoints from /.well-known/oauth-protected-resource and registers itself. Full wire-level flow in docs/auth.md.
Most of the surface is free. Four tools (challenges) plus the top-50 leaderboard rows (rows 11–50) require an active eighty Premium subscription, which is sold through the iOS App Store. See docs/premium-features.md for what's gated and how to upgrade.
Reference captures showing the server in use. Real demo data from staging; no PII.
Claude Desktop — claude-desktop/screenshots/
01-quests-browse-quests.png — browse_quests listing02-profile-character-sheet.png — get_profile (Atlas Lv 58)03-social-public-profile.png — view_public_profile rich response (Ember Lv 45)04-challenges-pending.png — list_challenges + get_challenge_detailsMCP Inspector — inspector/screenshots/
07-inspector-tool-list.png — tool catalog + connected state08-inspector-browse-quests.png — tool execution + JSON response09-inspector-view-public-profile.png — view_public_profile (Ember)10-inspector-prompts.png — the three MCP prompts.
├── README.md (you are here)
├── docs/
│ ├── architecture.md (thin proxy → backend handlers)
│ ├── tools-reference.md (all 35 tools with signatures)
│ ├── auth.md (OAuth 2.1 setup per client)
│ └── premium-features.md (what's gated, how to upgrade)
├── claude-desktop/
│ ├── claude_desktop_config.json (drop-in config snippet)
│ └── screenshots/ (4 + 2 bonus reference captures)
├── inspector/
│ └── screenshots/ (4 captures from MCP Inspector)
└── prompts/
├── README.md (index)
└── 01–08-*.md (8 example prompts)MIT.
Built by [@alemetlem](https://twitter.com/alemetlem). Star the repo if you want to see more public MCPs for consumer apps.
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.