Glossary
Key terms used across PromptPal product documentation.
- AI Playground
- In-app surface to run and compare prompt outputs across allowlisted models.
- Audit event
- Immutable record of an action (actor, actionType, org/workspace, metadata, timestamp).
- Clerk Directory Sync
- Production user provisioning path via Clerk—not the legacy PromptPal SCIM API.
- Company Brain
- Organization-scoped knowledge ingestion and template generation from connected tools—not Personal Brain.
- Effective policy
- Merged organization and workspace policy used by API enforcement middleware.
- Knowledge node
- A classified chunk of ingested org knowledge with handling AUTO_ACCEPT, REVIEW_QUEUE, or HOLD.
- Legal hold
- Policy flag that prevents retention jobs from deleting data during investigations.
- Lifecycle status
- Governance state of a prompt: draft, in_review, approved, or deprecated.
- Org permission
- Clerk-style permission key such as org:settings:manage or org:audit:view gating org APIs and sidebar items.
- Organization
- The enterprise tenant and control plane. Holds settings, identity, org policies, Company Brain, audit logs, and links to workspaces.
- Refine Agent
- Multi-turn conversational assistant for iterating on prompt text in saved sessions.
- Security scanner
- Risk analysis for prompt text—findings, scores, and optional handoff to Refine or Playground.
- Trust label
- Metadata on governed prompts indicating approval or provenance confidence.
- Workspace
- A collaboration boundary for prompts and members. May belong to an organization or exist as a standalone team workspace.
- Workspace role
- Membership role: owner, admin, editor, reviewer, or viewer—controls lifecycle and content actions.