AI Model Retirement Lint for hardcoded LLM model IDs

It reads the file you have open and flags every LLM model ID pinned in it that a vendor has already retired or has published a shutdown date for. Each finding gives the line number, the exact date, and the replacement model the vendor itself recommends. It ships 33 rules covering OpenAI, Anthropic and Gemini.

Engineers who ship OpenAI, Anthropic or Gemini calls with model IDs written into source, config or environment files. It is aimed at the person who owns the repository when a model ID goes dark in production - typically a backend or platform engineer on a team that added AI features during the last two years.

Because the chatbot's training data predates the retirement notice. The assistant that autocompleted claude-3-5-sonnet-20241022 into your config will confirm that ID is correct, since both come from the same period. A dated table checked against the vendor's own deprecation page does not have that blind spot.

Checking the open file is free and complete: all 33 rules run, every finding shows its date and replacement, and nothing is hidden or watermarked. The licence key covers a different job - scanning every file in the repository, failing a CI build, and rewriting pinned IDs to the vendor's replacement.

The extension is 29 US dollars once, with a 7-day full refund. The alternative is discovering a retired model ID when production starts returning errors, which is an unplanned incident plus the migration you were going to have to do anyway, done under time pressure rather than before the published date.