docs: add SVG detection test corpus#74
Merged
Merged
Conversation
2 benign controls + 5 malicious SVGs by AI-security category: DLP (sensitive data), prompt injection, system-prompt extraction, indirect injection + exfil, active-content/script-XSS. Payloads in title/desc/metadata/comment/hidden text/ CDATA script; each malicious file renders as a benign-looking image. Synthetic data, example.com URLs. Extends dlp-detection catalog + results matrix.
AIRS (profile "AI Gateway - Dev - Strict") skips DLP inspection entirely when input is classified as a valid SVG (svg root + SVG namespace) — same PII as plaintext is blocked. Injection/toxic/code detectors still fire on SVG. Adds svg-dlp-bypass.md (evidence, scan IDs, probe matrix, repro, recommendation) and updates the results matrix with measured outcomes.
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
1 participant
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Extends
docs/dlp-detection/with an SVG set — SVG is XML text, so it carries both DLP payloads and AI-prompt attacks.<title>/<desc>,<metadata>, XML comments, off-canvas/opacity-0<text>, CDATA<script>,javascript:hrefs,onerror):svg_mal_1_dlp— sensitive data (synthetic markers)svg_mal_2_prompt_injection— instruction overridesvg_mal_3_system_prompt— system-prompt extraction / jailbreaksvg_mal_4_exfil_injection— indirect injection + exfil toexample.comsvg_mal_5_script_xss— active content / scriptAdds
scripts/build_svg_corpus.py, extendscatalog.mdwith an SVG section, and adds SVG rows to the results matrix.Safety
example.com. Each malicious file verified to render as a harmless image viarsvg-convert.Test plan
mkdocs buildclean (no warnings/broken links)