Public documentation

Workbench Overview

Workbench is the shared workspace for teams that want local Roleplay results to become tracked security work.

It turns local CLI and CI results into:

Home action list
open findings
Evidence
attack-pack coverage
test history
protected agents
CI setup
API keys
privacy controls
ownership and status workflow

Current Release Behavior

Workbench stores workspace, project, run, finding, API-key, member, billing, and privacy settings for the shared workbench. Hosted production deployments use durable Postgres storage.

In local development only, workbench can use a local JSON store at:

.roleplay/cloud-db.json

You can point the app at another store with:

ROLEPLAY_CLOUD_STORE=/path/to/cloud-db.json

Production deployment assets, Postgres migrations, email/password workspace auth, billing endpoints, and readiness checks are included.

Main Screens

Screen	Purpose
Home	What needs attention across findings, fixes, evidence, and monitors.
Findings	Triage finding status, owner, fix state, and remediation.
Evidence	Step through the proof for a failed scenario.
Attack Packs	Review pack coverage and copy local CLI run guidance.
Scenarios	Build adversarial simulations with YAML preview.
Tests	Review local and CI uploads with commit context.
Agents	Manage protected agent targets.
Monitor	Create API keys, copy workflow snippets, set privacy controls, and schedule recurring attack checks.
Settings	Workspace members, API keys, retention, and project controls.
Billing	Builder and Team subscription state.

Workbench Plan Value

Use the included CLI to prove a failure locally. Use the workbench when that failure needs an owner, shared status, CI history, and regression tracking.