Prevent mitigation prompt injection by referencing mitigation item indexes#62
Open
Prevent mitigation prompt injection by referencing mitigation item indexes#62
Conversation
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Motivation
danger-full-accesssandbox.mitigateUX and the ability to select specific mitigation items.Description
buildMitigationPromptto accept a list of mitigation item indexes and build aFocusstring that references item numbers (e.g., "recorded mitigation item Spec v0.1: cstack workflow wrapper for Codex CLI #1"), rather than embedding raw mitigation text.startMitigationWorkflowto constructselectedActionsas{ index, action }entries and pass only the selected indexes intobuildMitigationPrompt, while still presenting the selected action text in the inspector output.--safe,--allow-dirty,--exec,--release,--issue) so mitigation workflows behave the same except that prompts no longer contain attacker-controlled text.recommendedActionsentry and asserts the spawned run summary references the mitigation item index (e.g., "recorded mitigation item Spec v0.1: cstack workflow wrapper for Codex CLI #1") and does not contain the malicious command text.src/inspector.ts,test/inspect.test.ts.Testing
npm test -- test/inspect.test.ts -t "can launch a mitigation workflow directly from a review inspection", and it passed.npm run typecheck(tsc --noEmit), and it passed.mitigate,mitigate <n>,mitigate <workflow>) functional while removing untrusted strings from spawned workflow prompts.Codex Task