@pretorian-worx/runclaudia-cli

The CLI for claudia — a diff-aware post-deploy test agent. Picks the relevant subset of your Playwright suite for a given diff, runs it against the deployed production URL, and reports the result to GitHub Actions, the merged PR, and Slack.

Install

npx -y @pretorian-worx/runclaudia-cli <subcommand> [args]

Or globally:

npm i -g @pretorian-worx/runclaudia-cli
claudia <subcommand> [args]

Subcommands

• claudia plan — diff-aware test plan generation (uses Anthropic)
• claudia map — build/inspect the route + endpoint + infra + DB + spec map
• claudia select — pick a diff-relevant subset of Playwright/Cypress specs
• claudia run — select + execute the subset against a deployed URL
• claudia ratings — aggregate 👍/👎 reactions across PR plan comments

claudia plan requires ANTHROPIC_API_KEY in the environment. All other subcommands are deterministic local code (zero API calls).

Post-deploy verification

This is the canonical use:

npx -y @pretorian-worx/runclaudia-cli run \
  --base $LAST_DEPLOY_SHA \
  --head $JUST_DEPLOYED_SHA \
  --target https://app.example.com \
  --slack-webhook $SLACK_WEBHOOK_URL

Result is auto-reported to:

• $GITHUB_STEP_SUMMARY when running in GitHub Actions
• A sticky comment on the merged PR that introduced the deployed commit
• Slack (if --slack-webhook or CLAUDIA_SLACK_WEBHOOK is set)

See the main repo README for the full workflow example and the larger product context.

License

MIT