DELARTINA
Opportunity discovery and manual research
Revenue Radar
Preparing Revenue Radar
Loading your saved Business Brain and latest opportunity deck.
Security Center
Revenue automation should prove its boundaries.
Delartina keeps the review workflow useful while live connections are absent. This page separates controls that work now from the production setup and legal review still required before public launch.
Current posture
Offline controls active
External sendingBlocked
OAuth token exchangeBlocked
Policy and proof gatesActive
Structured audit receiptsActive
Production database and authSetup required
Offline-ready does not mean publicly deployed. Provider credentials, production identity, durable storage, legal review, and incident operations remain explicit launch gates.
Control inventory
What the product enforces now
Each control names its purpose and the evidence a buyer or reviewer can inspect.
Browser and request boundary
Production CSP blocks framing and unsafe evaluation. Mutations enforce JSON content type, bounded bodies, trusted origins, actor context, and burst cooldowns.
CSP, frame, content-type, origin, forwarded-host, request-size, and rate-limit checks run in verification.
Outbound safety
No external message can leave this offline build. Draft, approval, proof, suppression, and workspace-policy gates remain visible and testable without provider credentials.
Blocked sends return a policy-safe error and keep the draft inside the workspace.
Role and workspace isolation
Owner, admin, and member capabilities are enforced at mutation routes. Local workspace identities remain isolated so one workspace cannot read another workspace's tasks or audit history.
Authorization and cross-workspace isolation are covered by API tests.
Auditable decisions
Account decisions, policy changes, plan changes, role edits, imports, drafts, blocked overrides, hierarchy changes, and suppression updates create structured receipts.
Evidence exports exclude credentials and retain before, after, reason, actor, and object context.
Data lifecycle
Clear enough to review without a security call.
The offline package exposes what data is useful, what is excluded, how retention is bounded, and where deletion still depends on production identity and storage.
Collect
Workspace records needed for account review, policy, drafts, approvals, tasks, team access, analytics, and audit evidence.
Minimize
Exports exclude provider tokens, OAuth secrets, API keys, cookies, credentials, and private payment identifiers.
Retain
Audit retention is explicit, plan-capped, and editable by workspace managers from governance settings.
Delete
Local sample data has an owner-confirmed reset path. Production deletion remains a launch requirement tied to durable identity and storage.
Threat model
Assume prompts, imports, browsers, and provider payloads can be hostile.
The repository threat model covers trust boundaries, protected assets, abuse cases, mitigations, residual risks, and the evidence required before a live provider path is enabled.
Protected assets
Workspace records, actor identity, policy, approval state, audit evidence, provider credentials, billing state, and outbound authority.
Primary threats
Cross-workspace access, forged origins, role escalation, unsafe sends, prompt-driven policy bypass, token leakage, oversized input, replay, burst abuse, and export leakage.
Current mitigations
Actor and role checks, origin and body guards, route-scoped rate limits, idempotency, policy gates, suppression records, blocked-action audit events, and secret/SAST scans.
Residual risks
Real OAuth, webhook verification, encrypted token storage, production auth, durable audit integrity, backup recovery, and legal operations must be validated in the deployed environment.
Launch boundary
The final review is operational, not cosmetic.
These checks remain required when the offline product is connected to real identity, storage, billing, AI, email, CRM, and Slack systems.
Production security review checklist
- Confirm production Postgres, backups, migration policy, and workspace isolation.
- Verify Clerk organization roles and session handling in the production domain.
- Complete OAuth provider review before accepting or storing any provider token.
- Validate token encryption, key rotation, Stripe webhook signatures, and secret ownership.
- Complete privacy, outbound-email, unsubscribe, retention, and incident-response review.
- Run the full verification suite and review dependency-license warnings before launch.
Inspect the evidence
Open the workspace control room.
Security, policy, permissions, suppression, export minimization, and red-team controls are visible in Settings without connecting a provider.