Draft
Conversation
Tailored system prompt for automated PR evaluation covering: - API surface fidelity (toString masking, descriptor shape, error types) - Prototype chain integrity and cross-frame leakage - DOM interaction safety and timing/race conditions - Platform-specific breakage patterns (WebView, Firefox xrays) - Third-party script compatibility (ad scripts, polyfill libraries) - Captured globals hygiene and missing captures - Messaging security (nativeData leakage, bridge trust boundary) - Prototype/scope attacks (document.__proto__, stack trace manipulation) - Data exfiltration vectors (postMessage, CSS injection, eval) - Configuration trust and iframe security - Risk classification aligned with cursor-review.yml workflow Co-authored-by: Jonathan Kingston <jonathanKingston@users.noreply.github.com>
Contributor
|
Cursor Agent can help with this pull request. Just |
Contributor
Build Branch
Static preview entry points
QR codes (mobile preview)
Integration commandsnpm (Android / Extension): Swift Package Manager (Apple): .package(url: "https://github.com/duckduckgo/content-scope-scripts.git", branch: "pr-releases/jkt/auto/injected-pr-evaluation-prompt-e4b5")git submodule (Windows): git -C submodules/content-scope-scripts fetch origin pr-releases/jkt/auto/injected-pr-evaluation-prompt-e4b5
git -C submodules/content-scope-scripts checkout origin/pr-releases/jkt/auto/injected-pr-evaluation-prompt-e4b5Pin to exact commitnpm (Android / Extension): Swift Package Manager (Apple): .package(url: "https://github.com/duckduckgo/content-scope-scripts.git", revision: "86a0e88a8ce1455ec3d27156a046ed5f61df3705")git submodule (Windows): git -C submodules/content-scope-scripts fetch origin pr-releases/jkt/auto/injected-pr-evaluation-prompt-e4b5
git -C submodules/content-scope-scripts checkout 86a0e88a8ce1455ec3d27156a046ed5f61df3705 |
Contributor
[Beta] Generated file diffTime updated: Mon, 09 Mar 2026 22:45:31 GMT |
![cursor[bot]](https://avatars.githubusercontent.com/in/1210556?s=60&v=4){kind=small}
Contributor
There was a problem hiding this comment.
Cursor Bugbot has reviewed your changes and found 1 potential issue.
Bugbot Autofix prepared a fix for the issue found in the latest run.
- ✅ Fixed: Prompt uses privacy-hostile term "user data"
- Replaced the heading text 'URL construction with user data' with 'URL construction with page-derived data' to align with DuckDuckGo's privacy-first language guidelines.
Or push these changes by commenting:
@cursor push 1a4276c6b9
Preview (1a4276c6b9)
diff --git a/.github/prompts/injected-pr-evaluation.md b/.github/prompts/injected-pr-evaluation.md
--- a/.github/prompts/injected-pr-evaluation.md
+++ b/.github/prompts/injected-pr-evaluation.md
@@ -111,7 +111,7 @@
### 4. Data Exfiltration Vectors
- **`postMessage` without origin restriction**: Any `window.postMessage()` or `iframe.contentWindow.postMessage()` without a specific `targetOrigin` (using `'*'`) leaks data to any listening frame.
-- **URL construction with user data**: Building URLs that include page-derived data (e.g., for image fetches, API calls) can leak browsing context to third parties. Verify all external URL construction.
+- **URL construction with page-derived data**: Building URLs that include page-derived data (e.g., for image fetches, API calls) can leak browsing context to third parties. Verify all external URL construction.
- **`eval()` / `Function()` / `new Function()`**: Direct code execution from strings. The codebase avoids this in C-S-S itself (broker-protection serializes functions for native execution). Flag any new dynamic code execution.
- **CSS injection**: If feature code injects `<style>` elements or sets `element.style` with page-derived values, CSS injection can exfiltrate data via `url()` or attribute selectors.
- **fetch / XHR from content script**: Content script fetches may carry cookies and bypass CORS. Any new network requests from feature code are high-risk.
Contributor
|
This PR requires a manual review and approval from a member of one of the following teams:
|
Contributor
There was a problem hiding this comment.
Web Compatibility Assessment
No findings.
.github/prompts/injected-pr-evaluation.md(entire file),severity: info: Documentation/prompt-only addition; no injected runtime behavior, API shims, prototype changes, or DOM mutations are introduced.
Security Assessment
No findings.
.github/prompts/injected-pr-evaluation.md(entire file),severity: info: No changes to trust boundaries (message-bridge, transports, captured globals, origin checks, or config execution paths); this cannot create a direct exploitable runtime path.
Risk Level
Low Risk — this PR introduces a static review prompt file only and does not alter shipped JavaScript execution in page or native messaging contexts.
Recommendations
- Add a CI check that validates required section headers in
.github/prompts/injected-pr-evaluation.mdto prevent prompt-format drift that could weaken automated review quality. - Add a short owner/maintenance note in the prompt (or adjacent docs) describing when architecture/security references should be updated, to reduce stale guidance risk over time.
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
2 participants
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.
Asana Task/Github Issue: N/A
Description
Adds a new system prompt (
.github/prompts/injected-pr-evaluation.md) tailored to evaluate Pull Requests for web compatibility and security issues within the Injected codebase and its messaging system.Testing Steps
Checklist
Please tick all that apply: