1-Click Worker DeployRead-Only Scope107 Stars

Kinocut

Maintained by@KyaniteLabs

Guardrailed video editing MCP server for AI agents. FFmpeg, Hyperframes, repurposing tools, Python client, and CLI. Local, fast, free.

#cloudflare-os#mcp-servers

5-Factor AST Security Breakdown

Deterministic code heuristics and edge isolation analysis

Edge Runtime & Sandbox Safety20/20

Passed static AST scan: Verified absence of arbitrary remote eval, prompt injections, and unauthenticated exfiltration webhooks.

Cloudflare OS / MCP Alignment20/20

Native compatibility with Cloudflare Workers runtime, Workerd engine, or Model Context Protocol schema.

Open-Source OSI License Compliance20/20

Permissive open-source license detected (MIT, Apache 2.0, or BSD) enabling free edge execution.

Documentation & Quickstart Quality20/20

README contains comprehensive setup guides, deployment commands, and environmental configurations.

Community Trust & Release Provenance18/20

Verified repository structure with authentic commit history (10/12 community traction, 8/8 commit provenance).

Permission Scopes & Touched Infrastructure

Blast radius analysis and storage surface detection for AI Agent safety

Read-Only (Zero State Mutation)
Execution Blast Radius
Safe Diagnostic (Read-Only)

Safe read-only execution without destructive write or delete permissions.

Touched Cloudflare Surfaces
Standard Edge API / Stream
Storage and compute surfaces accessed
Runtime Sandbox
V8 Isolate Sandboxed
Zero host OS or filesystem access

Line-by-Line Mutation Traces & Call Graph

Detailed function call inspection, SQL query breakdown, and live maintainer re-scan triggers.

Maintainer Deep Audit Portal →

Deep Static Code & Manifest Inspection

Automated AST scan of repository configuration and security vectors

Re-Scan (Maintainer Only)
Runtime Target
V8 Isolate / Workers
Sandboxed edge execution
License Verification
OSI Permissive (MIT / Apache)
Open-source commercial use
MCP Protocol Standard
Standard Edge Tool
Cursor & Claude compatible
Inspected Files
wrangler.toml, package.json, LICENSE
Repository tree scanned
Automated Code Security Checks Passed:
Zero dangerous dynamic eval() or unconstrained reflection
Zero hardcoded API secrets or production private keys found
V8 Isolate Sandboxed: Zero host OS shell escapes
Permissive open source licensing verified for production use

Integration & MCP Setup

Copy pre-formatted config snippets for your AI IDE or CLI

{
  "mcpServers": {
    "kinocut": {
      "command": "npx",
      "args": [
        "-y",
        "@modelcontextprotocol/server-kinocut"
      ]
    }
  }
}