Skip to content

0.4.2

Choose a tag to compare

View all tags
@dadachi dadachi released this 18 Oct 07:07
· 28 commits to main since this release
v0.4.2
4794738
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Each run reports what it cost (#167)

Every run now carries its own API spend, estimated at list price:

  • report.json has a cost block: estimatedUsd, calls, and per agent (planner / judge / repair) the calls, input / output / cache-write / cache-read tokens and USD.
  • The HTML report shows an "API cost (estimate)" table and the total in its header.
  • The CLI prints it after the verdict: cost: ≈ $0.2197 (25 API calls: planner $0.0273, judge $0.1924).
  • The MCP generate_app result includes it with the report, and the plugin's generate-app skill quotes it in its summary.

Pricing follows scripts/cost.sh (rates in src/models.ts; the repair agent's cost is the Agent SDK's own total). Before, a run's cost meant filtering the appended tmp/trace/ logs by time; on 2026-10-18 those per-call counts matched the Claude Console's usage export token for token. npm run cost still totals several runs from the traces.

The README's Runtime section also notes that --tests adds about 2.5 minutes and no API cost (the three suites run in parallel).

Tests

  • The smoke suite ignores NATIVEAPPTEMPLATE_MOBILE_IOS_DEVICE / _ANDROID_DEVICE exported in the developer's shell (#166).
Assets 2
Loading