docs: document Model Armor file and document scanning - #498
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
Collaborator
Author
|
bugbot run |
There was a problem hiding this comment.
✅ Bugbot reviewed your changes and found no new issues!
Comment @cursor review or bugbot run to trigger another review on this PR
Reviewed by Cursor Bugbot for commit 5694108. Configure here.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
BerriAI/litellm#31655 wired document and file attachment scanning into the Model Armor guardrail and it ships in v1.92.0. The behavior is fail closed, so inline documents that previously passed through unscanned can now be blocked with a 400, and a customer is already asking exactly how document links, oversized files, and masking behave with the fix
This adds a Document and File Scanning section to the Model Armor page covering which content blocks and document types are scanned, that a finding on a document always blocks even with mask_request_content enabled, and a table of the unscannable cases (file_id and remote URL references, documents over the 4 MB limit, more than 10 attachments per request, undecodable base64, scan errors) with the default fail-closed behavior versus fail_on_error: false. The fail_on_error param description now links to the new section
Note
Low Risk
Documentation-only changes with no runtime or config behavior modifications.
Overview
Documents v1.92.0 Model Armor behavior for inline document attachments on the Model Armor guardrail page.
Adds a Document and File Scanning section:
pre_call/during_callbyte API scanning for OpenAIfileand Anthropicdocumentblocks; supported document types vs pass-through (e.g. images); example payloads; HTTP 400 on policy matches; and that masking does not apply to documents. A follow-on subsection tables fail-closed cases (remote/file_idrefs, 4 MB limit, >10 attachments, bad base64, API errors) versus behavior whenfail_on_error: false, with a sample block error body.The
fail_on_errorparam blurb now calls out unscannable attachments and links to the new section.Reviewed by Cursor Bugbot for commit 5694108. Bugbot is set up for automated code reviews on this repo. Configure here.