Copilot Agent Mode

VI History Suite ships a built-in Model Context Protocol (MCP) server that exposes its LabVIEW VI comparison, history, and repository analysis to AI agents. When the extension is installed, Copilot agent mode discovers the server automatically, so there is nothing to configure.

The agent-mode surface requires VS Code 1.101 or later; on older editors the extension skips the registration and every other feature keeps working.

What You Can Ask

Work with your VIs in plain language, for example:

Summarize what changed in a VI comparison report.
Compare the last two revisions of a specific VI and explain the change.
Walk a VI's recent history and describe how it evolved.
Index the LabVIEW VIs in the repository, ranked by recent activity.
Review the LabVIEW VIs changed in a pull request and post a summary comment.
Fetch or validate the published VI-diff schemas.
Inspect the VI preview render cache (list, summarize, diagnose, or search entries).

Available Tools

The server provides 13 tools:

Tool	What it does	Runtime
`summarize_vi_comparison`	Concise what-changed narrative for a comparison report.	None
`get_vi_semantic_comparison`	Full semantic comparison model for a comparison report.	None
`compare_vi_revisions`	Runs a LabVIEW comparison between two Git revisions.	Comparison runtime
`summarize_vi_history`	Walks a VI's recent revisions and builds an evolution timeline.	Comparison runtime
`index_repository_vis`	Surveys tracked VIs, activity-ranked, with revision counts.	None (Git only)
`build_vi_pr_review`	Reviews a pull request by comparing every changed VI; the markdown form is a sticky PR comment.	Comparison runtime
`get_vi_semantic_schema`	Returns the published VI-diff JSON Schemas.	None
`validate_vi_semantic_document`	Validates a self-describing document against its schema.	None
`list_preview_cache`	Lists the entries in a VI-preview render cache directory.	None (read-only files)
`summarize_preview_cache`	Summarizes a preview cache directory: entry/byte counts, healthy vs flagged, and the flagged entries.	None (read-only files)
`diagnose_preview_cache`	Returns a preview-cache diagnostics snapshot so an agent can answer whether the cache is healthy in one call.	None (read-only files)
`search_preview_cache`	Finds cache entries by content marker (error, interactive, image, or empty); returns metadata only.	None (read-only files)
`get_preview_cache_entry`	Fetches one cache entry by key; returns metadata plus a file pointer, or the raw HTML on request.	None (read-only files)

Runtime Requirements

compare_vi_revisions, summarize_vi_history, and build_vi_pr_review run a real LabVIEW comparison, so they need a comparison runtime (the same host LabVIEW or Docker image you selected for the interactive compare flow) and may take a few minutes.
The remaining tools use Git, supplied data, or a local preview-cache directory only and return immediately.

Open VI-Diff Standard

The comparison, history, and repository-index results are published as versioned Draft-07 JSON Schemas so other tools can consume or validate them:

vi-history-suite/vi-semantic-comparison@v1
vi-history-suite/vi-semantic-history@v1
vi-history-suite/vi-repository-index@v1

Fetch a schema with get_vi_semantic_schema and validate any document that carries a schema field with validate_vi_semantic_document. The full tool reference is published in the project repository.