Setup

Project, Team, and Settings

Manage workspaces, roles, visibility, notifications, exports, governance, and privacy.

Workspace settings control the boundaries around decisions, signals, source access, notifications, exports, and team review.

Use these settings to decide where product judgement lives. Workspaces keep the judgement layer scoped to the people, products, agents, and North Star that should share the same context.

Workspaces (create, switch, types)

Ask The W supports multiple workspaces per organization. Use workspace boundaries for independent products, secure separation, or experiments.

Open /settings/workspaces to view, switch, or create workspaces.

Product reference

Product reference: Workspace settings with active organization, workspace list, and create workspace form

Find it in Ask The W at /settings/workspaces, or open Settings and choose Manage workspaces. Use this page to confirm the active organization, switch workspaces, create a private or shared workspace, and reset a workspace when needed.

Workspace types:

  • private: default for focused teams or single product streams.
  • shared: cross-team visibility while keeping shared boundaries.

Switching workspace changes which outcomes, decisions, signals, makers, connectors, notifications, and exports are active.

Creating a workspace requires owner or admin permission. A new workspace switches immediately after creation, so confirm the workspace switcher before connecting sources or making decisions.

Members and roles

Members belong to an organization and can access one or more workspaces. Owners and admins manage workspaces, membership, and governance.

Invite people only after visibility and review rules match the way the workspace will be used.

Use the least broad access that still lets the right people act. Workspace isolation prevents accidental context leakage across products, experiments, and operating groups.

Visibility levels

Use visibility levels to keep judgement layers clear:

  • Private: focused work for one person or a tight team.
  • Shared: cross-team visibility for product partners or leadership.

Choose the least broad visibility that still lets the right people act.

Agent decision review

Configure whether agent-origin decisions can auto-commit, queue for review, or require named approval.

Use stricter review for work that changes customer-facing behavior, security posture, billing, legal commitments, or irreversible data operations.

Agent governance should match trust level. Autonomous work can be helpful for low-risk docs and implementation trail capture, while high-impact changes should require explicit review.

Audit access

Audit trails preserve ownership and outcome changes. Org owners see the full audit trail. Members see entries on items they touched or can access inside their workspace.

Use audit access when reviewing ownership changes, connector changes, exports, agent approvals, or unexpected outcome edits.

The audit trail is meant to preserve the why behind corrections and workspace changes, not just the final state.

Notifications (categories, priority, delivery)

Notification preferences live at /settings/notifications.

Categories:

  • Approvals
  • Conflicts
  • Health changes
  • Refresh nudges

Each category can be set to High for immediate attention, Normal for digest-style review, or Off. Settings are workspace-scoped.

Export your data

Exports snapshot decisions, outcomes, and W context for governance or onboarding reuse.

Exports can include outcomes and decision links, counterfactuals when enabled, North Star progress, makers, and W conversation history.

Use exports before leadership reviews, governance handoffs, onboarding reuse, or migration rehearsals. Keep exported bundles scoped to the workspace that owns the operating context.

Synthesis defaults

Synthesis settings influence how often decisions refresh and how much counterfactual context appears in exported or generated views.

Tune these defaults when the workspace is either too noisy or too quiet. A high-signal source with clear owners can tolerate more automation. Sparse or sensitive work should prefer stricter review and smaller exports.

Privacy: what stays local, what is sent

The plugin stores signals and decisions on your machine. Nothing leaves until you connect a workspace.

When you connect a workspace, compact summaries, signal excerpts, file paths, and commands can be sent. Full file contents are not sent by default. Outcomes live in the Ask The W app.

The plugin redacts common secrets before capture, including OPENAI_API_KEY=, OpenAI-style keys, bearer tokens, AWS access keys, JWTs, emails, and other operational secrets. Redaction is on by default and can be configured in ~/.askthew/config.json:

{
  "redaction": {
    "enabled": true
  }
}

Avoid placing raw secrets in prompts, signal excerpts, or command text even with redaction enabled.

Before workspace connection, local identity and pending captures stay on your machine in ~/.askthew. After workspace connection, compact source material can be sent to Ask The W so the app can connect signals, decisions, and outcomes.