Helixar Security — Claude MCP Connector

Three agentic-AI security tools, exposed as a remote MCP server, listed in the Claude Connectors Directory.

Tool	What it does
`helixar_inspect_mcp`	Scan an MCP server (URL or raw manifest JSON) against Sentinel detection rules. Returns risk score, findings, and a Claude-generated security brief. Quick mode is free + authless (top 8 rules). Deep mode runs all 26 rules with an API key.
`helixar_hdp_validate`	Validate an HDP delegation chain against IETF draft `draft-helixar-hdp-agentic-delegation-00`. Surfaces scope escalations, depth violations, expired hops, missing signatures. Every output cites the IETF draft + Zenodo DOI.
`helixar_triage_alert`	Triage a Vigil / ATP detection payload into a kill-chain stage (Preparation / Positioning / Expansion / Objective) with a Claude-generated narrative in your choice of executive, technical, or brief format.

Quick start

npm install
npm test
npm run build
npm start          # stdio MCP server

Add to Claude (custom connector)

The hosted server lives at https://mcp.helixar.ai. To use it before Anthropic lists it in the directory:

Open Claude → Settings → Connectors → Add custom connector
URL: https://mcp.helixar.ai
Auth: OAuth 2.0 (Claude handles the flow)
Save and refresh — the three tools appear in the tool picker.

For local development, point Claude Desktop at node /path/to/helixar-mcp/dist/server.js as a stdio server.

Architecture

Language: TypeScript ESM (Node 20+)
MCP SDK: @modelcontextprotocol/sdk (official Anthropic)
Validation: Zod for tool input schemas
Narration: Anthropic SDK with deterministic fallback when no API key is configured
Hosting: Cloudflare Workers (src/worker.ts, deployed to mcp.helixar.ai)
Auth: OAuth 2.0 + Dynamic Client Registration (required for directory listing)

Auth tiers

Mode	Auth	Tools / scope	Purpose
Quick / public	none	`inspect_mcp` (top-8 rules), `hdp_validate`	Maximum reach — zero-friction for community adoption
Authenticated	API key (OAuth2)	`inspect_mcp` deep mode (26 rules), `triage_alert`	Pilot customers + paid tier

Repository layout

src/
├── server.ts              # MCP stdio entrypoint
├── worker.ts              # Cloudflare Workers HTTP adapter (Phase 7)
├── lib/
│   ├── narrate.ts         # Anthropic call + deterministic fallback
│   ├── sentinel-rules.ts  # 26 Sentinel detection rules (top-8 quick + 18 deep)
│   ├── hdp-schema.ts      # HDP chain types + 9 validation rules
│   └── vigil-parser.ts    # Vigil/ATP payload normaliser + stage classifier
└── tools/
    ├── inspect-mcp.ts     # helixar_inspect_mcp implementation
    ├── hdp-validate.ts    # helixar_hdp_validate implementation
    └── triage-alert.ts    # helixar_triage_alert implementation
tests/
└── (mirrors src/)

IP protection

Per the implementation plan §6, internal detection methodology, Hunch Mode internals (IOB pipeline, weighted signals, anomaly scoring), Vigil sensor implementation, and exact thresholds are never exposed in this codebase. Public surface is rule IDs, severity buckets, public-safe detection categories, and remediation guidance only. CI includes a forbidden-symbol scan in lib/vigil-parser.ts to catch accidental leaks.

License

Apache-2.0 — see LICENSE and NOTICE.

Helixar Security — Claude MCP Connector