Prevent cross-user data leakage by salting remote cache keys #2362
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
Introduces support for user-specific salting in cache key generation. This enables higher-level cache isolation mechanisms to incorporate user-derived entropy while keeping the cache entry interface backward compatible.
Fixes a security issue where remote cache entries could be fetched using predictable cache keys. Remote cache fetch and store operations now use cache keys salted with a SHA256-derived value based on the user's EXPECTED_PARROT_API_KEY, ensuring that cached responses are isolated per user. Local cache access remains unsalted to preserve backward compatibility.
Updates pricing estimation logic to use the same salted cache key strategy as remote cache operations. This ensures accurate cache hit predictions and consistent behavior between pricing estimation and runtime execution.
Adds a test that validates cache key salting behavior for remote cache operations. The test confirms that identical prompts from different users produce distinct remote cache keys while preserving backward compatibility for unsalted local cache lookups.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
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Outdated suggestions cannot be applied.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.
Problem
Remote cache keys were derived from predictable hashes of prompts and model
parameters. Any authenticated user could query arbitrary cache keys,
potentially retrieving cached responses belonging to other users.
This created a privacy and security risk when prompts contain proprietary
or sensitive data.
Solution
This PR introduces user-specific cache key salting for all remote cache
operations:
hash derived from the user's EXPECTED_PARROT_API_KEY
Impact
Files Changed
cache_entry.pycache.pyjobs_pricing_estimation.pytests/test_cache_salting.py