Repository navigation
RL Environment for Indirect Prompt Injection - #1051
Merged
Merged
Conversation
Signed-off-by: Makesh Sreedhar <makeshn@nvidia.com>
…rouping, logging Signed-off-by: Makesh Sreedhar <makeshn@nvidia.com>
…rouping, logging Signed-off-by: Makesh Sreedhar <makeshn@nvidia.com>
…erifier edge cases Signed-off-by: Makesh Sreedhar <makeshn@oci-hsg-cs-001-vscode-01.cm.cluster>
makeshn
marked this pull request as draft
April 10, 2026 21:03
added 2 commits
April 10, 2026 14:11
Signed-off-by: Makesh Sreedhar <makeshn@oci-hsg-cs-001-vscode-01.cm.cluster>
makeshn
marked this pull request as ready for review
April 10, 2026 21:19
Signed-off-by: Brian Yu <bxyu@nvidia.com>
bxyu-nvidia
previously approved these changes
Apr 16, 2026
cmunley1
reviewed
Apr 16, 2026
Signed-off-by: Makesh Sreedhar <makeshn@oci-hsg-cs-001-vscode-01.cm.cluster>
Contributor
|
it seems the data validation test is failing |
Contributor
i believe it is just missing example_metrics.jsonl -- probably due to .gitignore |
Signed-off-by: Makesh Sreedhar <makeshn@oci-hsg-cs-001-vscode-01.cm.cluster>
Contributor
Author
|
@cmunley1 Fixed it |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Add Indirect Prompt Injection (IPI) Environment
Summary
Introduces a new security environment for training and evaluating LLM resistance to indirect prompt injection attacks. The environment simulates realistic tool-use scenarios across 9 enterprise domains where injected instructions are hidden within tool-returned data (e.g., patient chart notes, email bodies, application notes), testing whether the agent follows the injected instruction or correctly ignores it.
Domains
Healthcare, HR/Recruiting, Ecommerce, Education, Insurance, IT Helpdesk, Legal, Logistics, Real Estate. Each domain has 10-12 simulated tool handlers.
Verifier: Scans the agent's output trace for evidence that the injected instruction was followed.
Reward