Platform guide: Software lifecycle check

This method runs a standalone tool on the person's own machine. No platform table or field is named here, because none was confirmed against a source for this method. Where a source would be needed, it says Not found.

Where the inventory export comes from

The tool reads a CSV with a name column and a version column. For exports it knows, it offers --profile flexera, --profile servicenow-mbs, and --profile rapid7. Unverified: no one has run any of these profiles on a real export. For any other export, map the columns with --columns FILE; the tool ships templates/columns.example.json as a pattern. Which report or table in a given platform produces the right columns: Not found.

Running the tool, by assistant

• Claude Code and Codex: run the commands on the person's machine, next to the inventory.
• Microsoft Copilot: cannot run local tools. The person runs each command themselves and pastes back only the summary count lines. Never paste the inventory, a report section, or a line that lists item IDs into Copilot.
• Claude app and ChatGPT: not supported by this method. Running it there would mean uploading the inventory to a remote sandbox.

Optional counts-only server

The tool also ships an optional counts-only server for desktop chat apps, and a Claude Desktop package that is unsigned. See Mcp setup at the pinned commit. The tool's docs say each was tested once on one Mac. Unverified: anything beyond that. This method runs the command line, not the server.

What the tool does not replace

The tool reports state from its bundled data. It does not decide what to replace, what risk to accept, or whether a KEV match affects a version. Those are the owners' calls.

Sources checked

Nothing was fetched for this guide. The tool facts come from the software-lifecycle tool's own documents at the pinned commit, v0.7.0, 1ebc996e52b728507a4113aee6c67ea0e208a9e7, as summarized in the skill this method was adapted from.