swenyai/SWEny E2E — Agentic Browser Tests

Run agentic browser tests against a deployed app. AI drives the browser, no test scripts.

View on GitHub

Trust Signals

Scorecard Score
not yet scored
Maintenance Recency
Stale
License
None
namedescriptionrequireddefault
workflowPath to the E2E workflow YAML file to execute (relative to working-directory).yes
claude-oauth-tokenClaude Code OAuth token. Use this OR anthropic-api-key. OAuth tokens do not consume API credit.no
anthropic-api-keyAnthropic API key for Claude. Use this OR claude-oauth-token.no
base-urlBase URL of the deployed app under test. Exposed to the workflow as BASE_URL.no
cli-versionVersion of @sweny-ai/core to install (the package that ships the `sweny` CLI). Defaults to latest.nolatest
agent-browser-versionVersion of agent-browser to install. Defaults to latest.nolatest
node-versionNode.js version to install.no24
working-directoryWorking directory to run the workflow from. Defaults to the repository root.no.
screenshots-pathPath (relative to working-directory) where the workflow writes screenshots and artifacts. Uploaded after the run.noresults/
artifact-nameName of the uploaded screenshots artifact. The job's run id is appended for uniqueness.noe2e-screenshots
artifact-retention-daysHow long to retain the screenshots artifact (days).no14
verboseWhen 'true', print each tool call's input and output inline. Truncated for readability; locally, the CLI's --stream flag yields full untruncated NDJSON. Helpful for debugging a workflow that halts or routes unexpectedly. Requires @sweny-ai/core >= 0.1.94.nofalse

no outputs