CLI Reference

Global Options

wendell --version
wendell --help

Run Wendell commands from your application repository root. Commands that read configuration load ./wendell.toml by default. Use --config <path> only when the config lives somewhere else or you are selecting an alternate file.

wendell register

Register a local runner and store credentials.

wendell register

Options:

Option	Description
--api-url	Override the Wendell API URL. Production is the default.
--email	Wendell account email.
--agent	Agent display name for this runner.
--provider	Model or runtime provider.
--model	Model or runtime name.
--computer-name	Runner machine name.
--password-stdin	Read password from stdin.

wendell login

Store an existing API key.

printf "%s\n" "$WENDELL_INKPASS_API_KEY" | wendell login \
  --api-key-stdin \
  --validate

wendell whoami

Show the active identity.

wendell whoami

wendell auth export

Export the stored runner credential for CI secret setup.

wendell auth export --format raw | gh secret set WENDELL_INKPASS_API_KEY

Formats:

Format	Use
env	Print an export WENDELL_INKPASS_API_KEY=... line for local shell setup.
raw	Print only the key for piping into secret managers.
json	Print the key plus profile metadata for setup scripts.

The command reads the local credential created by wendell register or wendell login. It does not create a new key.

wendell doctor

Check local or CI setup before running a hosted suite.

wendell doctor

Use --validate to verify the active runner credential against the Wendell API:

wendell doctor --validate

wendell doctor checks the config file, credential source, API URL, and agent_command parsing without printing secrets. With --json, it emits a machine-readable pass/fail report for CI setup scripts.

wendell playbook create

Create a Playbook draft.

wendell playbook create \
  --name "Refund Agent Regression" \
  --workflow-summary "A support agent evaluates refunds." \
  --source ./refund-playbook.md \
  --project-ref refund-agent \
  --domain customer-support-refunds \
  --extract

Options:

Option	Description
--name	Required. Playbook or test suite name.
--workflow-summary	Required. Short workflow description.
--source	Source file to ingest. Can be repeated.
--source-text	Inline source material.
--project-ref	Customer or project reference.
--org-ref	Organization reference used when auth is unscoped.
--created-by	Creator label. Defaults to wendell-cli.
--domain	Workflow domain.
--extract	Immediately generate the first Playbook summary.
--json	Print JSON output.

wendell playbook review

Review extracted Playbook primitives and questions.

wendell playbook review wdraft_abc123

wendell playbook apply

Apply a review patch.

wendell playbook apply wdraft_abc123 \
  --file ./review.json

wendell playbook approve

Approve the latest Playbook summary.

wendell playbook approve wdraft_abc123 \
  --reviewer "you@example.com" \
  --generate-suite

wendell suites list

List accessible published suites.

wendell suites list

wendell suites generate

Generate a suite candidate from an approved draft.

wendell suites generate \
  --draft wdraft_abc123

wendell suites publish

Publish a generated suite candidate.

wendell suites publish \
  --draft wdraft_abc123

wendell suites show

Show a published suite that your runner can access.

wendell suites show refund-agent-regression

wendell suites configure

Write wendell.toml and an adapter template for a hosted suite.

wendell suites configure \
  --suite refund-agent-regression \
  --project refund-agent

Options:

Option	Description
--suite	Required. Published test suite slug.
--project	Project slug to write into wendell.toml. Defaults to the suite slug.
--config	Config path to write. Defaults to ./wendell.toml.
--adapter	Adapter template path to write. Defaults to scripts/wendell_agent_adapter.py.
--agent-command	Agent command to write into wendell.toml. Defaults to python <adapter>.
--force	Overwrite existing config or adapter files.

wendell run

Run a local agent against a published suite.

wendell run \
  --suite refund-agent-regression

Options:

Option	Description
--suite	Required. Published test suite slug.
--config	Override config path. Defaults to ./wendell.toml in the current project directory.
--api-url	Override the Wendell API URL. Production is the default.
--json	Print JSON output.

wendell run evaluates configured gates after uploading the run. In advisory mode, failed gates are printed and the command exits successfully. In blocking mode, failed gates return a nonzero exit code for CI.

With --json, the remote object includes private_report_url when Wendell can create a private dashboard handoff link.

wendell runs watch

Fetch the current hosted status for a run.

wendell runs watch run_abc123

When the run came from CI, the human-readable output includes the source repository, workflow, CI run URL, revision, and runner version. Use --json for the full run envelope.

wendell runs report

Fetch the private hosted report for a run.

wendell runs report run_abc123

With --json, the response includes private_report_url when Wendell can create a private dashboard handoff link.