Privacy Policy
Last updated: October 24, 2025. This privacy policy outlines how AiCorner handles developer data, API credentials, and telemetry across our registry.
Introduction
AiCorner ("we," "our," or "us") operates the premier index and registry for AI agent tools, Model Context Protocol (MCP) servers, developer APIs, and agent skills. We are committed to maintaining the trust of developers, enterprise engineering teams, and open-source contributors who rely on our platform.
This Privacy Policy applies to all services, developer portals, API endpoints, and command-line interfaces operated by AiCorner. By accessing or using our platform, you acknowledge the collection and handling practices described herein.
Data Collection
We collect information strictly necessary to provide a high-performance directory and facilitate deterministic tool calling for autonomous agents.
Usage & Telemetry
- API invocation patterns and latency metrics
- Skill search queries and filter selections
- Workflow performance metrics and error rates
- Anonymized IP addresses and device user agents
API Keys & Credentials
- Encrypted API authentication tokens
- OAuth 2.0 access tokens and client scopes
- Developer profile and organization metadata
- Publisher verification records
How We Use Data
Data collected across our platform is utilized exclusively to enhance system reliability, maintain registry accuracy, and protect developer sandboxes:
- Agentic Workflow Optimization: We analyze skill execution patterns to optimize parameter binding and reduce context window latency.
- Registry & Category Accuracy: Real-time metadata guides dynamic category tool counts and verifies active provider availability.
- Proactive Security & Anomaly Detection: Telemetry monitors for malicious prompt injections, rate-limit violations, and unverified endpoints.
Third-party Services & Registered Integrations
AiCorner interfaces with external Model Context Protocol (MCP) servers, agent skills, developer APIs, and compute providers to deliver real-time tool discovery and execution capabilities. AiCorner does not own or operate third-party remote endpoints; developers interacting with external servers are subject to the independent privacy terms of each respective provider.
Registered Model Context Protocol (MCP) Servers
The following MCP servers represent real protocol adapters cataloged in our registry. Each server exposes deterministic tools, resources, and prompt templates over Stdio or SSE transport channels.
| MCP Server Name | Transport Layer | Capabilities & Scope | Official Source / Repository |
|---|
Registered Agent Skills & Knowledge Blueprints
Agent Skills are local or published instruction packages that inject domain-specific workflows and parameter constraints into AI agents.
| Agent Skill Name | Category | CLI Installation Command | Author / Maintainer |
|---|
Compute Providers & Infrastructure Ecosystem
The infrastructure layer supporting AiCorner query resolution, static site generation, and telemetry storage adheres to strict zero-retention data privacy standards.
| Provider / Platform | Service Category | Data Exchanged | Compliance & Security |
|---|---|---|---|
| Anthropic Claude API | Model Inference & MCP Protocol | Prompt context & JSON tool call payloads | Zero Data Retention (ZDR) |
| OpenAI API | LLM Inference & Function Calling | Function call arguments & completions | SOC2 Type II / TLS 1.3 |
| GitHub Enterprise | Version Control & OAuth 2.0 | Repository metadata & user profiles | OAuth 2.0 Scoped Credentials |
| AWS / Google Cloud | Database & Storage Infrastructure | Encrypted registry records & asset logos | AES-256 Encryption at Rest |
Data Security
We implement robust technical and organizational security controls to protect developer assets against unauthorized access, loss, or manipulation:
- AES-256 Encryption: All database records and secrets are encrypted at rest using AES-256, and data in transit is encrypted via TLS 1.3.
- Magic Bytes Asset Validation: Every uploaded logo and asset file is verified against binary magic byte headers before storage to prevent malicious payload uploads.
- Zero-Trust Boundary Control: Server-side API endpoints validate input parameters with strict Zod schemas to block SQL injection and cross-site scripting.
User Rights & Data Control
Developers have full autonomy over their published tools, skills, and account metadata. You may request data export, credential revocation, or listing removal at any time by contacting our compliance team at privacy@aicorner.us or via our Contact Page.