docs(config): document image allow-list for llm output - #416
Open
KyryloKorotych wants to merge 1 commit into
Open
KyryloKorotych wants to merge 1 commit into
KyryloKorotych wants to merge 1 commit into
Conversation
Add a Security & Compliance page covering the domain allow-list that gates every image rendered from LLM/assistant output, and document the allowedImageDomains customer config item. Generated with AI Co-Authored-By: codemie-ai <[email protected]>
This branch has not been deployed
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Summary
Documents the image domain allow-list delivered in EPMCDME-12950 (codemie!3976, codemie-ui!1616) — the security control that gates every image rendered from LLM/assistant output against a per-deployment hostname list, closing a prompt-injection exfiltration channel.
Adds a full guide under Security & Compliance and documents the
allowedImageDomainscustomer config item in the configuration reference.Changes
docs/admin/security/llm-output-image-allow-list.md— threat model, scope of the gate, the 9-condition decision order, entry format, configuration, rollout and verification stepscustomer-feature-configuration.md— new CONTENT SECURITY group with theallowedImageDomainsrow in the component overview table, a dedicated section under Component Categories, and an entry in the full configuration examplesidebars.tsand cross-linked it from the Security & Compliance index; freedsidebar_position: 3by moving Network Policies to4exfiltrationandsrcsetto the cspell dictionaryTesting
npm startnpm run build(zero broken links)Quality Checks
npm run checkpasses (typecheck + lint + commitlint)<text>must be`<text>`)static/img/) — not applicable, no images addedAdditional Notes
Verification performed:
npm run checkandnpm run buildboth exit0, with zero broken links reported. A visual pass withnpm startwas not run — the checkboxes above reflect only what was actually verified.Source-of-truth discrepancy worth flagging to the reviewer: the description on codemie-ui!1616 documents configuration via the
VITE_ALLOWED_IMAGE_DOMAINSfrontend env var, but that approach is absent from the merged diff — there are no.env,config.js, ordeploy-templateschanges. The implementation readsappInfoStore.getAllowedImageDomains(), sourced fromGET /v1/config→allowedImageDomains, which matches codemie!3976. This documentation follows the code, not the !1616 description — that description should be updated so it no longer contradicts both.Implementation details documented here that appear in neither MR description, taken from the diff:
markedrenderer override — so raw HTML in model output is gated identically to markdown image syntaxsrc,srcset,poster,background;srcsetcandidates are checked individually<source>element is removed rather than badged, letting the<img>fallback apply; on other elements only the offending attribute is stripped