LIFE.SAVER RULES

Policy rules UI · Active Platform Conversion · v0.8.3
← Dashboard Settings Memory Actions Notifications Functional Audit Super Admin Launch Prep Login AUTH CHECKING WIZARD SAFE
Logged In As
—
—
Phase
7.10
Rules permission QA.
Policy Engine
READY
Phase 6.10 tests complete.
Execution
DISABLED
No executors or external writes.
Founder Autonomy Control Center

Policy Rules UI

Phase 7.10 keeps the safe seven-step rule wizard plus Content, Support, Ads, Global Caps, Rule Simulation preview, Pause Integration Display, and Policy Audit UI, then adds dedicated Rules Permission QA. The founder can clearly see when master pause or category pauses override automation. This phase does not create database policies, edit backend policy rows, auto-run anything, publish content, send support replies, change budgets, pause campaigns, or write to external platforms.

Default ask Pause first Hard caps Most restrictive wins No execution Local preview save Content form Support form Ads form Caps form Simulation preview Pause warning Audit UI Permission QA
Phase 7.9

Pause Integration Display

Shows when global or category pause is overriding rule previews and future automation.

CHECKING
Pause override status Checking pause status…

If pause status cannot be loaded, LIFE.SAVER should behave conservatively and keep future automation in manual review.

  • Global pause: checking
  • Content pause: checking
  • Support pause: checking
  • Ads pause: checking

Pause override rule: global pause, category pause, and emergency safe mode must beat auto-approval and future executors. Example: “Auto-run is currently paused globally.”

Use Refresh Pause Status to re-check the existing autonomy status endpoint. This display is read-only and does not pause or resume anything itself.

Section 1 · Existing Pause Controls

Master Pause

Shows the current autonomy pause state and links back to Settings for the existing pause controls.

CHECKING
Global Pause—
Content—
Support—
Ads—

Master pause remains the highest-priority safety control. If active, future auto-approval and future executor execution must stay blocked.

Open Pause Controls
Phase 7.2 retained

Create Rule Wizard

Build a policy-rule preview through seven safe steps. Save is local-preview only until backend policy editing and permissions are added in a later approved phase.

NO DB WRITE
  1. 1. Action type
  2. 2. Platform/channel
  3. 3. Condition
  4. 4. Decision
  5. 5. Caps
  6. 6. Review
  7. 7. Save
Step 1 — Choose action type

Choose which future action family the policy would apply to. This does not enable the action.

Step 2 — Choose platform/channel

Scope the policy to a safe platform/channel. Empty values are avoided so the future rule shape is predictable.

Step 3 — Add condition

Choose one Phase 6.4 condition operator. More complex condition builders come later.

Tip: for risk_below, use low, medium, high, or critical. For amount/confidence operators, use a number.

Step 4 — Choose decision

Auto-approve in this wizard is only a future policy decision preview. It does not trigger execution in this package.

Step 5 — Set caps

Caps are included in the preview JSON using the Phase 6.6 cap names. Empty values are omitted.

Step 6 — Review

Review the policy shape before saving the local preview draft. This JSON is not sent to the backend in Phase 7.2.

{}
—
Step 7 — Save

Safe save behaviour: Phase 7.2/7.3 saves browser-local previews only. It does not insert into the policies table, does not call policy editing endpoints, and does not enable auto-run.

Not saved yet.
Step 1 of 7
Section 2 · Phase 7.3 retained

Content Rules

LOCAL PREVIEW

Content-specific policy controls for platform, approved style, max posts/day, allowed content type, auto-publish preference, and approval risk threshold.

Not saved yet. 
{}

Safe Phase 7.3 behaviour: this form creates a local preview only. It does not create a policy row, does not enable auto-publish, and does not publish content.

Section 3 · Phase 7.4 retained

Support Rules

LOCAL PREVIEW

Support-specific policy controls for ticket category, confidence threshold, auto-reply preference, escalation categories, reply cap, and sensitive-ticket exclusions.

Not saved yet. 
{}

Safe Phase 7.4 retained behaviour: this form creates a local preview only. It does not create a policy row, does not send support replies, and does not connect to Gmail/helpdesk APIs.

Section 4 · Phase 7.5 retained

Ads Rules

LOCAL PREVIEW

Ads-specific policy controls for platform, campaign scope, max daily budget change, max percentage change, always-ask threshold, and pause rules.

Not saved yet. 
{}

Safe Phase 7.5 behaviour: this form creates a local preview only. It does not create a policy row, does not connect to Meta/Google/TikTok/Snapchat/Pinterest/Microsoft Ads, does not change budgets, and does not pause campaigns.

Section 5 · Phase 7.6 retained

Global Caps

Caps settings UI for posts per day, auto-replies per day, ad budget change per day, model token/cost usage, and actions per hour.

LOCAL PREVIEW
Posts/daymax_posts_per_dayBlocks content auto-approval when exceeded.
Auto-replies/daymax_support_auto_replies_per_dayBlocks support auto-reply overuse.
Ad budget change/daymax_ad_spend_change_per_dayProtects real money before ad phases.
Model token/cost usagemax_model_tokens_per_day + max_model_cost_per_dayProtects Claude/model usage budget.
Actions/hourmax_actions_per_hourPrevents action floods.
Not saved yet. 
{}

Safe Phase 7.6 behaviour: this form creates a local preview only. It does not create or update policy rows, does not save caps to the database, does not enable auto-run, and does not execute actions.

Section 6 · Phase 7.7 retained

Rule Simulation Preview

Test a proposed action against the current local rule previews before enabling anything. The result explains whether the action would be auto-approved, ask/manual review, or blocked.

LOCAL SIMULATION
Simulation result This proposed Instagram post would be: ASK ASK

Run the preview to see the rule reasoning.

Not saved yet. 
{}

Safe Phase 7.7/7.8/7.9/7.10 behaviour: simulation runs in the browser against local previews only. It does not create a policy row, does not enable a rule, does not call a policy write endpoint, does not run executors, and does not write to external platforms.

Phase 7.9 · Section 7

Rule History / Policy Audit

Shows local-preview audit metadata for rule changes and recent simulated policy decisions before database-backed policy history is enabled.

LOCAL AUDIT
Created by—
Updated by—
Disabled by—
Last triggered—
Recent decisions—
Rule / preview Created by Updated by Disabled by Last triggered Recent decisions
Loading local policy audit preview…
Not saved yet. 
{}

Safe Phase 7.9/7.10 behaviour: this Policy Audit UI reads browser-local previews and saved simulation JSON only. It does not query policy audit tables, does not write audit records, does not enable rules, and does not execute anything.

Phase 7.10 · Permission QA

Rules Permission QA

Confirms only authorized workspace roles can edit rule previews before any future database-backed policy editing endpoints are enabled.

QA READY
Effective rule-edit access Checking role permission…

Owner can edit. Admin can edit only when rule-edit access is allowed. Viewer and unauthorized sessions cannot edit.

  • Owner: can edit
  • Admin: can edit if allowed
  • Viewer: cannot edit
  • Unauthorized request blocked: no token or wrong workspace cannot edit
Detected/current role—
Owner edit test—
Admin edit test—
Viewer/unauthorized—
Scenario Role Admin allowed Authorized request Expected result Reason
Loading rules permission QA preview…
Not saved yet. 
{}

Safe Phase 7.10 behaviour: this Rules Permission QA checks UI gating and expected role decisions only. It does not create or update policy rows, does not save permissions to the database, does not add policy editing endpoints, and does not execute anything.

Phase 15.6 retained · Mobile readable

Proactive Trigger Registry

Mobile-readable registry for ROAS drop, new support ticket, scheduled content slot, weekly ad review, pending action reminder, and failed executor event. This is display and preview only.

FRAMEWORK ONLY
No proactive job runs from this screen.

Triggers may identify a future route or specialist, but this mobile UI does not create actions, send notifications, call Claude, invoke tools, call connectors, execute, or auto-run.

ROAS dropManual review route

Flag meaningful blended ROAS declines for founder review.

New support ticketSupport specialist route

Surface new tickets without sending a reply automatically.

Scheduled content slotContent specialist route

Remind founder of planned content windows without publishing.

Weekly ad reviewAds specialist route

Queue a safe weekly review summary without changing budgets.

Pending action reminderApproval queue route

Remind founder about waiting approvals without approving them.

Failed executor eventEscalation route

Keep failed executor events visible and never hidden.

Static mobile registry loaded. 
{}

Phase 7.10 Safety Boundary

Allowed in this phase:
Seven-step rule creation wizard, content rules form, support rules form, ads rules form, caps settings form, rule simulation preview, pause-state display, policy audit UI, local preview JSON, browser-local preview save, navigation, section layout, and documentation/testing for rules permission QA.

Not allowed in this phase:
No backend policy creation endpoint, no database policy insert/update, no audit database save, no permission database save, no cap database save, no simulation database save, no ad connector, no ad budget update, no campaign pause, no auto-run toggles, no sandbox executor, no real executor, no content publishing, no support sending, and no external platform writes.