Local run & debug · Verentis CLI

verentis dev run executes a local script with a real scoped token and injected context.

It does not enable the Python SDK's offline local-filesystem mode. For an offline script, call engine.configure_local(files_root=..., runtime_dir=..., setting_aliases=...) in a separate local-only runner and execute python run-local.py. Keep that call out of the production script you package for Verentis. Use dev run when the script must exercise real workspace files and platform APIs.

Run

verentis dev run transform.py --workspace <id> --params '{"message":"hi"}'
verentis dev run transform.py --arg --verbose --env LOG_LEVEL=debug

<script>

Required

The local script to run.

--file <path>

The workspace file path the run pretends to target (default: /<script name>).

--tool <tool>

Engine tool name (default: execute).

--mode <mode>

Execution mode (default: request-response).

--arg <value...>

Arguments passed to the script. Repeatable.

--env <KEY=value...>

Extra environment variables. Repeatable.

--params <json>

Tool parameters JSON.

--scopes <scope...>

Explicit token scopes (default: manifest spec.permissions).

--manifest <path>

Manifest to read spec.permissions from.

--boundary <path>

Confine the token to a directory subtree, like platform runs.

--python <interpreter>

Python interpreter (default: python3).

--keep-context

Keep the generated .verentis/context.json after the run.

Debug

--debug waits for your debugger before the first line runs — it wraps the script with debugpy --listen <port> --wait-for-client:

verentis dev run transform.py --debug          # debugpy on :5678, waits for a client
verentis dev run transform.py --debug 5679     # custom port

Scaffold VS Code launch configs (an attach and a launch entry):

verentis dev vscode                            # writes .vscode/launch.json
verentis dev vscode --port 5679

Attach flow

Start dev run … --debug, then in VS Code pick the generated attach configuration to connect to debugpy on the chosen port.