Add trajectory importer for agent-evaluation conversation traces - #5
Merged
Conversation
Reads recorded agent-eval result files (system/user/assistant/tool messages, an injections map, security/utility fields, a schema common to agent-eval harnesses) and emits neutralized PAB Trace and GroundTruth pairs. Maps roles to channels, locates each injection to set true origin and surface span, computes hop counts honestly (0-1 hop is expected, never fabricated), and redacts model, pipeline, and project identifiers by default. Authorization is derived through the same is_authorized path as the generator and harvester, so all corpus sources share one authorization logic; surprising derivations are logged for review rather than forced. New pab import-trajectories command. Episodes that cannot be safely attributed are skipped with reasons. Every emitted trace passes the origin-leak guard. 569 tests, floor gate green.
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
Reads recorded agent-eval result files (system/user/assistant/tool messages, an injections map, security/utility fields, a schema common to agent-eval harnesses) and emits neutralized PAB Trace and GroundTruth pairs. Maps roles to channels, locates each injection to set true origin and surface span, computes hop counts honestly (0-1 hop is expected, never fabricated), and redacts model, pipeline, and project identifiers by default. Authorization is derived through the same is_authorized path as the generator and harvester, so all corpus sources share one authorization logic; surprising derivations are logged for review rather than forced. New pab import-trajectories command. Episodes that cannot be safely attributed are skipped with reasons. Every emitted trace passes the origin-leak guard. 569 tests, floor gate green.