Governance
Approvals
The approval matrix and risk-routing layer. High-impact actions are gated behind a human decision, scored for risk and routed to the right approver before anything goes public or spends budget.
These actions always require a human gate before an agent can execute them:
Publish PostLaunch CampaignSpend BudgetSend MessageSend BroadcastCreate Public ProfilePublish AI Realistic Content
Pending
3
awaiting decision
High / critical
3
risk level
Approved
1
cleared
Rejected
1
blocked
Pending Queue
Risk-gated actions waiting on a human approver
@lumi.sleep — 3AM pain reel
Publish PostSocialOpsAgent
Risk score35/100
AI-assisted creative — disclosure label attached.
CW002 Meta launch — €2,200
Launch CampaignAdsManagerAgent
Risk score70/100
Spend > €1k and AI-realistic creative present.
Create YouTube channel @lumisleeplab
Create Public ProfileSocialOpsAgent
Risk score40/100
Decision Log
3 resolved approvals
| Object | Action | Requested by | Approver | Risk | Status | Decided |
|---|---|---|---|---|---|---|
| AI-influencer reel (Sofia Calma) | Publish AI Realistic Content | AIInfluencerAgent | — | High | Changes Requested | — |
| Telegram broadcast — 1,840 members | Send Broadcast | CommunityAgent | user_founder | Medium | Approved | 2 days ago |
| TikTok BOFU launch | Launch Campaign | AdsManagerAgent | — | Critical | Rejected | 3 days ago |