Semantic Shield — AI Skill Safety Validation API
Published by Simply Semantics (https://www.simplysemantics.com)

================================================================================
OVERVIEW
================================================================================

Semantic Shield is a security-first SaaS component that protects AI agents,
bots, and users from unsafe skills, plugins, and MCP tools.

Every skill in the Semantic Shield registry is vetted by US-based security
experts with 30+ years of enterprise experience — including work for
US Homeland Security. No AI-only reviews. 100% REAL human experts.

Agents and developers query any skill's safety profile via API or MCP and
receive a safety score (0-100), risk level, threat details, and a clear
install/reject recommendation — all in real time, at machine speed.

Unvetted skills can be submitted for expert evaluation with full lifecycle
tracking. Continuous 0-day threat monitoring keeps all assessments current.

================================================================================
KEY CAPABILITIES
================================================================================

* Real-Time Safety Lookups — Query any skill's trust profile via API
* 100% Human-Reviewed — US-based experts with 30+ years security experience
* Safety Scoring 0-100 — Risk levels: critical, high, medium, low, safe
* Threat Detection — Prompt injection, data exfiltration, over-privileged
  permissions, malicious code, privacy violations, and more
* Continuous Monitoring — 0-day vulnerability tracking keeps scores current
* Expert Evaluation — Submit unvetted skills for review with tracked requests
* Webhook Alerts — Pro+ tiers get notified when threats affect vetted skills
* Free Search — Find skills by name or provider at no quota cost
* Agent-First Design — API-key auth for bots, agents, and autonomous workflows
* MCP-Compatible — Works with OpenClaw, LangChain, and other AI frameworks

================================================================================
USE CASES
================================================================================

- AI Agent Builders: Verify skills before autonomous installation
- Enterprise Security Teams: Enforce trust policies on agent ecosystems
- Bot Developers: Ensure plugins and tools meet safety standards
- SaaS Platforms: Integrate trust checks into skill marketplaces
- AI Consultants: Advise clients on safe AI tool adoption
- Compliance Officers: Maintain auditable records of vetted tools

================================================================================
HOW IT WORKS (Quick Start)
================================================================================

1. Search — Find skills by name or provider (free, no quota cost)
2. Check — Get full trust details: safety score, risk level, threats,
   recommendation (costs 1 lookup from monthly quota)
3. Submit — If not vetted, submit for expert evaluation (costs 1 inquiry)
4. Decide — Use the safety score and recommendation to install or reject
5. Monitor — Vetted skills are continuously re-checked for new threats

================================================================================
API ENDPOINTS
================================================================================

All endpoints require an API key via x-api-key header.

GET  /api/v1/search?q=...&provider=...
     Search for skills (free, no quota cost)
     Returns: skill_id + provider pairs

GET  /api/v1/check?skill_id=...&provider=...
     Full trust details (costs 1 lookup)
     Returns: safety_score, risk_level, threats, recommendation

POST /api/v1/validate
     Submit skill for expert evaluation (costs 1 inquiry)
     Body: { skill_id, provider, skill_url }
     Returns: request_id, status

================================================================================
SAFETY SCORING
================================================================================

  Score     Risk Level    Action
  -------   ----------    ----------------------------
  90-100    safe          Install / proceed
  75-89     low           Install, note minor risks
  50-74     medium        Warn user, ask confirmation
  25-49     high          Recommend against
  0-24      critical      Block installation
  null      unknown       Not vetted — submit or warn

================================================================================
PRICING
================================================================================

Free Tier .............. $0/month
  25 lookups/month, 5 evaluation inquiries/month
  Basic trust verdict, FAQs support

Pro Tier ............... $29/month
  100 lookups/month, 20 evaluation inquiries/month
  Priority evaluations, full reports, webhook alerts, email support

Enterprise ............. Custom pricing
  Unlimited lookups and inquiries
  Private queue, SLA, dedicated support, custom integrations

================================================================================
CONTACT & SUPPORT
================================================================================

Website:  https://www.simplysemantics.com
Product:  https://www.simplysemantics.com/semantic-shield.html
Email:    info@simplysemantics.com

License:  MIT (for any open-source components)

