Migrated Enterprise GPT plugin loses its files, and no one promises they'll arrive
A forum post says a custom GPT converted to a plugin in ChatGPT Enterprise runs without its supporting files, which breaks tasks that rely on code in those files.
The short version
- One user migrated a custom GPT to a plugin in an Enterprise ChatGPT account.
- The plugin relies on instruction files and templated Python code that produce an HTML artefact.
- The plugin files are not transferred to the chat by default, so only the skill file runs.
- The poster says the same plugin works fine in personal ChatGPT.
- The poster asks for migration instructions; we saw no OpenAI reply.
Why it matters
If your team builds workflows on custom GPTs, a migration can quietly change what the tool has access to. A plugin that runs but lacks its code files can fail or give weak output without an obvious error. Test migrated tools with real inputs before anyone depends on them.
What the user reports
As community.openai.com showed in a forum thread, a user converted a custom GPT into a plugin inside an Enterprise ChatGPT account and hit a problem. The tool depends on several files: instructions plus templated Python code (code with blanks filled in for each run). Together they are meant to produce an output such as an HTML artefact, a self-contained web page.
According to the poster, the plugin files are not carried into the chat by default. The plugin ends up running only from its skill file, which makes the whole exercise futile in their words.
Why a missing file matters
Code can only run if it can find what it needs. Here the instructions and Python templates are the working parts, so a plugin that sees only the skill file has the recipe but not the ingredients.
The thread's title describes this as a missing guaranteed file-mounting contract. Mounting means making files available in the place where the code runs. A contract would be a firm promise about when and where they appear. The post itself does not explain how OpenAI handles this internally, so treat that reading as a plain-language gloss on the title.
Personal versus Enterprise
The most striking detail is the contrast. The poster says the same kind of plugin works without trouble in personal ChatGPT. That suggests the difference lies in the Enterprise setup or the migration path, though the post does not say which.
The user ends by asking whether there are migration instructions to follow when converting GPTs to plugins. The post does not mention an answer from OpenAI, and we found none.
How to read this
This is a single user's account on a community forum. It has not been confirmed by OpenAI, and it may reflect a setup choice or a gap in documentation rather than a defect.
Still, it shows a practical risk for anyone moving custom GPTs into a managed workplace account: the tool can look intact after migration while missing parts it needs.
What to watch
- Will OpenAI publish migration guidance for converting custom GPTs to plugins in Enterprise?
- Will OpenAI confirm whether plugin files are meant to be available in Enterprise chats by default?
Try it
If you have migrated a GPT to a plugin, run it once with a realistic task and check that the output actually used your instruction and code files.
Sources and links
Liked it? Say why
Add a line for other readers.
No account needed. We never show your email. Comments with links, ads or abuse are held and not shown. Your comment shows after a quick check.
Listen to the AI Wire headlines
Listen to the AI WireListen to our daily AI shows every morning: on your commute, in the shower, wherever you are
Listen to the daily showsFirst reported by community.openai.com · Read the original. Written by the SI Agent News Desk from the reporting above.Written with AI from the source above. Every number is checked against the source. Spot a mistake? Corrections.