chore: release

[primer]

This PR was opened by the Changesets release GitHub action. When you're ready to do a release, you can merge this and the packages will be published to npm automatically. If you're not ready to do a release yet, that's fine, whenever you add more changesets to main, this PR will be updated.

Releases

@primer/agent-eval@0.13.0

Minor Changes

• 72742af: Require runner on parsed experiment and benchmark trial outputs. Legacy output
  files without a runner default to copilot-cli when parsed. Consumers constructing
  trial output objects directly must include runner: 'copilot-cli' or
  runner: 'copilot-sdk'.

• a5d1340: Add support for running benchmark experiments with the benchmark option.

  Calls to getExperiment and listExperiments now require benchmarksDirectory,
  including when loading only scenario-backed experiments.

Patch Changes

• 922afff: Fix output-token totals for Copilot CLI and SDK trials by collecting usage
  statistics instead of relying on message events that no longer include token
  counts. Preserve compatibility with older session output.
• 7fd247f: Validate manifest consistency through schemas before loading configurations or
  trial artifacts. Reject metadata IDs that differ from their manifest keys,
  duplicate benchmark capability scenario IDs, and conflicting metadata across
  benchmark result shards. Duplicate trial IDs and other merge conflicts now report
  schema errors with field paths.