Skip to content

Add Garak tool#44

Merged
marcorosa merged 39 commits intodevelopfrom
tool/garak
Aug 21, 2025
Merged

Add Garak tool#44
marcorosa merged 39 commits intodevelopfrom
tool/garak

Conversation

@marcorosa
Copy link
Copy Markdown
Member

@marcorosa marcorosa commented May 6, 2025

This PR adds garak as a tool that can be run from STARS.
Not all Garak's probes are supported by STARS, as many of them fall outside our working scenario. Thus, only the ones related to prompt injection, jailbreak, system prompt leak and insecure code generation have been integrated.
This is to say, the following:

  • dan
  • encoding
  • goodside
  • latentinjection
  • malwaregen
  • phrasing
  • promptinject
  • suffix

Still TODOs:

@marcorosa
Copy link
Copy Markdown
Member Author

Double check attack score as computed here before merging the PR

@marcorosa
Copy link
Copy Markdown
Member Author

Langchain upgrade do v3 work in progress in a dedicated branch. To be merged here.

@marcorosa
Copy link
Copy Markdown
Member Author

langchain v3 is supported (and the corresponding branch might be merged into this one), but #23 is needed, too (introducing other conflicts).

@marcorosa
Copy link
Copy Markdown
Member Author

pyrit update wip in #52

@marcorosa marcorosa marked this pull request as ready for review June 20, 2025 14:49
@marcorosa marcorosa requested a review from a team as a code owner June 20, 2025 14:49
@marcorosa
Copy link
Copy Markdown
Member Author

MERGE AFTER #46

@marcorosa marcorosa mentioned this pull request Jun 24, 2025
@marcorosa
Copy link
Copy Markdown
Member Author

Fix conflicts and garak attacks' return json format

@marcorosa
Copy link
Copy Markdown
Member Author

TODO BEFORE MERGE:

  • make AttackResult return class consistent to other attacks (see Caroline's work)
  • save results to db

@marcorosa marcorosa changed the title Add Garak tool [DO NOT MERGE] Add Garak tool Aug 18, 2025
@marcorosa marcorosa changed the title [DO NOT MERGE] Add Garak tool Add Garak tool Aug 21, 2025
@marcorosa marcorosa merged commit fcc885c into develop Aug 21, 2025
6 checks passed
@marcorosa marcorosa deleted the tool/garak branch September 12, 2025 16:45
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant