# Super Agent Skill
> The MCP infrastructure layer for AI agents. Connect any agent to a live registry of skills, playbooks, souls and guardrails — then let the proprietary forge research, author, evaluate and continuously evolve them.

Site: https://superagentskill.com
MCP endpoint: https://superagentskill.com/api/mcp
Transport: MCP Streamable HTTP (JSON-RPC 2.0). Required header: `Accept: application/json, text/event-stream`
Long-form agent manual: https://superagentskill.com/agents.md
Sitemap: https://superagentskill.com/sitemap.xml

## What it is (TL;DR)
Super Agent Skill is to AI agents what npm is to Node and what Twilio is to telecom: a single
connection that gives any MCP-compatible agent (Claude, Cursor, Codex, Grok, custom) instant
access to thousands of installable, version-controlled, continuously-evolving capabilities.
No retraining. No SDK. No DevOps. One MCP URL → one sentence → a specialist.

## Who it's for
- AI engineers wiring agents into production
- Teams that already use Claude/Cursor/Codex/Grok and want them to actually do the job
- Domain experts (doctors, lawyers, marketers, fintech ops) monetizing their expertise as packages

## Quickstart (30 seconds)
1. Add MCP server in your agent's config:
   ```
   {"mcpServers":{"superagentskill":{"url":"https://superagentskill.com/api/mcp"}}}
   ```
2. Authorize via OAuth 2.1 + PKCE (browser opens automatically).
3. Ask the agent: "make me a cardiologist" — it discovers, installs and evolves.

## Primitives (mental model)
| type      | what it is                                              | use it for                          |
|-----------|---------------------------------------------------------|-------------------------------------|
| skill     | a single capability                                     | ECG read, MEDDPICC, KYC analysis    |
| playbook  | a multi-step workflow                                   | chest-pain triage, sales motion     |
| soul      | tone / personality / decision style                     | Challenger rep, calm clinician      |
| guardrail | safety boundary, runs before output                     | no advice without citation, no PII  |

## Tools (full reference)
- list_packages({type?, query?, limit=20}) — list registry items
- search_registry({query, limit=10}) — semantic search by name/description
- get_package({slug}) — full manifest (system_prompt, rules, examples, compatibility)
- request_primitive({type, brief, industry?}) — auto-create a missing primitive
- get_skill_trust({slug}) — trust score, downloads, ratings, drift signals

## Decision rules for agents
1. Detect 1–3 domain keywords from the user turn.
2. `search_registry({ query })`. If no hit, `request_primitive`.
3. Adopt the returned `system_prompt` + `rules` for the turn.
4. Stack at most: 1 soul, 1–N skills, 0–1 playbook, 0–N guardrails.
5. Guardrails always run last and may refuse.

## Pricing
- Hacker — $0/mo · 3 installs, registry read, MCP gateway
- Agent Pass — $19/agent/mo · unlimited upgrades, SkillForge AI, weekly reports
- Enterprise — custom · private registry, SAML SSO, RBAC, audit logs
Creators keep 80–85% of every package sold.

## Key links
- Connect:        https://superagentskill.com/connect
- Marketplace:    https://superagentskill.com/marketplace
- Pricing:        https://superagentskill.com/pricing
- Docs:           https://superagentskill.com/docs
- MCP docs:       https://superagentskill.com/docs/mcp
- Discover:       https://superagentskill.com/discover
- Forge:          https://superagentskill.com/forge
- Agent manual:   https://superagentskill.com/agents.md