-
Notifications
You must be signed in to change notification settings - Fork 18
Submission: AgentBusters - copyrightcat #19
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
base: main
Are you sure you want to change the base?
Submission: AgentBusters - copyrightcat #19
Conversation
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Pull request overview
This PR adds a "CopyrightCat" security arena submission from team AgentBusters that tests AI resistance to jailbreak attacks attempting to bypass copyright restrictions. The scenario uses puzzle-solving techniques (letter-by-letter extraction, word games, encoding challenges) to indirectly extract copyrighted song lyrics.
Changes:
- Added CopyrightCat scenario plugin with puzzle-based attack templates and LLM-based copyright detection
- Added test results demonstrating successful attack in round 2 and baseline pass
- Added template-based and scenario-specific attacker implementations
Reviewed changes
Copilot reviewed 8 out of 8 changed files in this pull request and generated 8 comments.
Show a summary per file
| File | Description |
|---|---|
| test_results/result.json | Test execution results showing attacker success in round 2 |
| test_results/baseline_passed.json | Baseline test results confirming defender works correctly with benign inputs |
| test_results/attack_succeeded.json | Attack success confirmation with metadata |
| template_attacker.py | Template-based attacker using fixed puzzle-solving prompts |
| scenario_copyrightcat.toml | Configuration file defining endpoints and scenario parameters |
| plugin.py | Scenario plugin implementing attack/defense logic and copyright detection |
| copyrightcat_attacker.py | Scenario-specific attacker with 15 attack templates |
| README.md | Documentation explaining attack vectors and implementation |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Demo Video
https://youtu.be/D991loZgFS4
Direct Logic Override (Jailbreak) - Evaluating AI resistance to creative prompt manipulation techniques such as puzzle-solving attacks to bypass copyright restrictions and fictional framing to extract harmful content