Skip to content

feat(jvchat): replay the selected Debug Interactions tick - #223

Merged
eldonm merged 3 commits into
devfrom
feat/debug-retest-tick-replay
Sep 13, 2026
Merged

eldonm merged 3 commits into
devfrom
feat/debug-retest-tick-replay

Conversation

@Tharickv75

Copy link
Copy Markdown
Contributor

Summary

Enables tick replay for Debug Interactions by passing earlier-turn assistant and tool messages directly to model re-runs.

  • Adds messages, tool_choice, and parallel_tool_calls parameter support to query_model_action.
  • Dynamically constructs full chat histories (system, history, user, and replay) in DebugInteractions.tsx.
  • Adds editable This-turn tool replay (JSON) and Tools (JSON) panels for re-testing model calls.
  • Resolves missing recorded tool schemas using stub reconstruction based on available tool names.

Related issues

N/A

Type of change

  • Bug fix (non-breaking)
  • New feature (non-breaking)
  • Breaking change
  • Documentation
  • Chore / tooling / CI

Checklist

  • I read CONTRIBUTING.md and (for subsystem work) the local CLAUDE.md.
  • pre-commit run --all-files passes.
  • pytest tests/ passes; I added/updated tests for new behavior.
  • Bug fixes cite file:line in the description.
  • Docs / CHANGELOG.md updated where relevant (entry under [Unreleased]).
  • No ADR was modified in place (changes of decision are written as a new, superseding ADR).

Notes for reviewers

  • Includes unit tests covering observationReplayForMetric and resolveRetestTools in debugToolCalls.test.ts.

@Tharickv75 Tharickv75 self-assigned this Sep 11, 2026
@Tharickv75
Tharickv75 requested a review from eldonm as a code owner September 11, 2026 20:33
Eldon Marks and others added 2 commits September 13, 2026 17:47
Concurrent waiters hydrate distinct Python instances after serialized
create; object-identity checks flake under full-suite pre-push.

Co-authored-by: Cursor <[email protected]>
Preserve assistant response text in observation replay, stub tools from
this tick's calls instead of inheriting an unmatched sibling surface, and
fingerprint replay sync so log refreshes reset editors.

Co-authored-by: Cursor <[email protected]>
@eldonm
eldonm merged commit a4c7408 into dev Sep 13, 2026
3 checks passed
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.

2 participants