| github-token | GitHub token used to post or update the PR report. | yes | — |
| mode | analyze, analyze-and-run, or run. | no | analyze |
| app-url | Explicit preview, staging, or localhost URL for browser execution. | no | — |
| target | Named app target from .preflight-scout/config.yml app.targets. | no | — |
| target-env | Exact local/staging URL from config, or automatic URL discovery when set to auto. | no | auto |
| detect-deployment-url | With target-env auto, use the latest successful GitHub Deployment URL when app-url is not set. | no | true |
| mission-id | Run one generated browser mission by id. | no | — |
| all-candidates | Run every generated browser mission sequentially. | no | false |
| mission-limit | Number of LLM-ranked automation candidates to run when all-candidates is false. | no | — |
| max-turns | Maximum LLM browser turns per mission. | no | — |
| storage-state | Playwright storageState JSON to load before browser missions. | no | — |
| save-storage-state | Path where Playwright storageState JSON should be written after browser missions. | no | — |
| trace | Capture Playwright trace.zip for each browser mission. | no | true |
| headless | Run Playwright in headless mode. | no | true |
| output-dir | Directory for report, JSON, HTML, screenshots, and run artifacts. Defaults under RUNNER_TEMP. | no | — |
| comment | Post or update a PR comment with the Markdown report. | no | true |
| upload-artifact | Upload the report bundle as a GitHub Actions artifact. | no | true |
| artifact-name | Custom GitHub Actions artifact name. | no | — |
| fail-on | CI gate behavior. Use never, needs_attention, or failed_only. | no | needs_attention |