AUTONOMY PERMISSION REGISTRY

TheoB needs explicit permissions before autonomy becomes action.

A permission registry for auto-allowed, review-required, and blocked action classes across dashboards, memory, signals, workflows, publishing, compute, secrets, and irreversible operations.

◎CivilizationFull system map◉WatchObserve live◇PredictStrategic foresight▶ExecuteAction routes✦AcademyLearn + train▣VaultMemory + proof◌VillageMissions + exchange⌁MobilePocket command
Universal Time Scroll

Every TheoB pathway can move through Past, Present, and Future without losing context.

Present

Read current signals, conditions, and live context.

World SignalsWatchSystem Health
Saved discoveries · event and discovery dates

Discoveries through time

Living Intelligence Ecosystem · Saved evidence shared across Library, capsules, world and time.

Connecting this view to its project…

🎙
Universal Voice Orb

Voice ready

SYSTEM STATE
RepresentedFounder ControlledHuman Reviewed
ACTION PERMISSIONS

Autonomy without permission is just a fancy way to make trouble faster.

Autonomy Permission Registry defines what TheoB may do automatically, what requires review, and what must be blocked before action routes into workflows, publishing, compute, credentials, memory, or irreversible operations.

Autonomy OSAgent PermissionsHuman Override
Auto-Allowed
dashboard.enrich.auto

Update internal dashboards, labels, summaries, and non-sensitive status views.

RiskLow public impact, low reversibility risk.
SafeguardNo public publishing, no external sending, no irreversible changes.
Auto-Allowed
memory.note.auto

Store low-risk operational notes, audit summaries, and non-sensitive system events.

RiskMemory bloat or stale context.
SafeguardMemory retention gate, trust decay, and correction pathway.
Auto-Allowed
signal.classify.auto

Classify incoming signals by source, domain, urgency, and likely routing path.

RiskWrong classification may create routing noise.
SafeguardNo autonomous high-impact action from classification alone.
Auto-Allowed
information.free.read.auto

Research, retrieve, compare, and summarize free public information through fixed, terms-reviewed sources within enforced request caps.

RiskSource error, rate-limit abuse, privacy leakage, or mistaken confidence.
SafeguardRead-only access, source receipts, timestamps, provenance, confidence, rate caps, backoff, and kill switches.
Delegated-Allowed
information.recurring.bounded.delegated

Qualified agents or internal teams may authorize bounded recurring free-source observation within an existing provider contract.

RiskRunaway polling, provider pressure, stale collection, or unnoticed scope expansion.
SafeguardDelegated approval receipt, cadence cap, volume cap, backoff, kill switch, review window, and automatic suspension on anomaly.
Review-Required
information.paid.once.review

Use paid, gated, credentialed, or quota-sensitive information access after one scope-specific authorization.

RiskSpend, contractual obligations, credential exposure, or unexpected downstream impact.
SafeguardSingle authorization receipt with provider, purpose, cap, expiry, and stop condition; warn and reconfirm when impact appears materially larger than understood.
Protected-Review
information.critical.escalate

Escalate information actions with critical, irreversible, safety-sensitive, legal, privacy, or infrastructure impact.

RiskIrreversible or materially misunderstood consequences.
SafeguardExplain likely ramifications before approval; require explicit protected review and preserve rollback where possible.
Review-Required
workflow.prepare.review

Prepare workflow drafts, deployment plans, legal drafts, public claims, or sourcing recommendations.

RiskMedium impact if executed without approval.
SafeguardWorkflow Execution Gate and human review before execution.
Review-Required
public.publish.review

Publish public claims, product language, research interpretations, ceremony guidance, or market-facing updates.

RiskReputation, legal, cultural, or claims exposure.
SafeguardReality verification, claims safety, legal review, and human approval.
Review-Required
spend.compute.review

Run recurring, expensive, or high-volume compute workflows.

RiskCost spike or runaway loop.
SafeguardEconomic Pressure Gate, stop conditions, and founder approval.
Blocked
credential.exposure.blocked

Expose, print, store publicly, or transmit raw secrets, tokens, passwords, or API keys.

RiskSevere security compromise.
SafeguardBlock immediately and route to Key Master without exposing secret values.
Blocked
irreversible.action.blocked

Delete critical assets, send binding agreements, perform payments, alter legal status, or execute high-impact actions without approval.

RiskIrreversible operational, legal, financial, or trust damage.
SafeguardBlock unless explicit human approval and correct gate path exist.