Skip to content

Spike scripts

Throwaway research scripts used to validate Graph API shapes or third-party tooling before writing production code. Kept in-tree so future contributors can see what was probed and why.

These are not part of the supported surface. They may break on Graph schema changes. Treat as documentation.

spike_a365_work_iq.py

Inspect local Agent 365 Work IQ setup. Prints the configured Work IQ server metadata from ToolingManifest.json and validates that Word is present before implementation begins.

Usage

python scripts/spike_a365_work_iq.py

Set ENTRACLAW_A365_TOOLING_MANIFEST to override the manifest path. Defaults: ToolingManifest.json, .a365/ToolingManifest.json.

What it does

  • Loads the manifest.
  • Lists registered Work IQ MCP servers.
  • Confirms the Word tool surface is registered.

Used to derisk the A365 Work IQ MCP integration before any wiring in production code. See docs/platform-learnings/microsoft-agent-365.md for the platform context.

spike_file_comments.py

Capture verbatim Graph beta JSON shape for file comments.

Usage

python scripts/spike_file_comments.py <share-url>

What it does

  • Mints a fresh Agent User token (picks up just-PATCHed Files / Sites scopes).
  • Resolves the SharePoint URL via GET /shares/{share-id}/driveItem.
  • Dumps raw responses from:
  • GET /beta/drives/{drive}/items/{item}/comments
  • GET /beta/drives/{drive}/items/{item}/comments/{comment_id}/replies (only if at least one comment exists)

Used to pin down the exact JSON shape before implementing the file-comment MCP tools. The findings are captured in docs/architecture/PLAN-files-mcp-tools.md.