Montage is an automated readiness assessor and scaffolding engine for the agent ecosystem. It inspects existing codebases, decomposes them into discrete functional capabilities across five structural layers, and answers whether an application can, and should, be integrated into autonomous agent workflows.

Key Features

  • 5-Layer Capability Extraction: Scans REST routes, hierarchical CLI command trees, utility scripts, authored prompt personas, and domain engine packages.
  • 5-Dimension Readiness Scorecard: Evaluates codebases on Interface Automation, Data Interchange, Execution Safety, Auth and Config, and Runtime Packaging.
  • Spectrum-Aware Scaffolding: Synthesizes Pure Skills for advisory procedures, Skills with Code for token-saving data filtering, and persona-clustered MCP servers.
  • Agent Plugin Bundling: Exports turnkey packages conforming to the Agent Plugins Specification (plugin.json, mcp.json, and specialized skills/).
  • Interactive Web Dashboard: Features a self-contained Lit WebComponent dashboard deployable in a single command to Google Cloud Run.

Explore the source code on GitHub: github.com/ghchinoy/montage.