You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
As of this issue being written, this candidate was recorded in the retired local task-candidate backlog and is being migrated to GitHub Issues as the task-candidate source of truth. No Agent Eval Lab task bundle, reference artifact, reference verification artifact, or Codex smoke trial exists for it yet.
Why interesting: compact JavaScript library task with behavior and tests.
Risks / checks: no linked upstream issue was found during initial curation; clarify intended behavior from the PR body and tests before promoting.
Why change
The starter suite needs more varied realistic tasks before broader capability reporting. This candidate would add compact JavaScript API behavior coverage.
What to build
Triage whether this candidate should become a publishable task. If it is promoted, the eventual implementation issue should create a task bundle with pinned repo and commit, deterministic graders, a verified reference artifact, generated task card, and one fair Codex smoke trial.
Acceptance criteria
Decide whether to promote, defer, or reject this candidate.
If promoting, ensure the issue has enough detail to become ready-for-agent: expected behavior, grader strategy, setup constraints, and stop conditions for suspicious or resource-heavy setup.
If deferring or rejecting, comment with the reason.
Parent
Related to #10.
Current behavior / design
As of this issue being written, this candidate was recorded in the retired local task-candidate backlog and is being migrated to GitHub Issues as the task-candidate source of truth. No Agent Eval Lab task bundle, reference artifact, reference verification artifact, or Codex smoke trial exists for it yet.
Candidate notes
8cb53ea5c3329032a1db47be019b717d8350fb0e285e19dc3b6c53099f196396ffc7e769fb9555a5Why change
The starter suite needs more varied realistic tasks before broader capability reporting. This candidate would add compact JavaScript API behavior coverage.
What to build
Triage whether this candidate should become a publishable task. If it is promoted, the eventual implementation issue should create a task bundle with pinned repo and commit, deterministic graders, a verified reference artifact, generated task card, and one fair Codex smoke trial.
Acceptance criteria
ready-for-agent: expected behavior, grader strategy, setup constraints, and stop conditions for suspicious or resource-heavy setup.Blocked by
None - can start immediately.