Capabilities Pillars Use Cases Domain Packs Governance Industry Context Docs Get in touch →
Governance

The envelope around every run

Every run is tenant-scoped, sandboxed, budget-limited, and traced. The 16-type runtime envelope is the complete set of hard constraints that apply to every agent, regardless of domain pack.

The envelope

Four control surfaces

The 16 runtime types group into four surfaces. Each surface is enforced by the runtime and cannot be bypassed by an agent.

Scope

  • Declared in manifest
  • Enforced per call
  • No silent widening
  • Policy-referenced

Tool mediation

  • Allow-listed per agent
  • Invoked via runtime
  • Arguments validated
  • All calls traced

Budget

  • Runtime budget per call
  • Economic budget per pack
  • Tracked separately
  • Overruns halt the run

Isolation

  • Tenant-scoped
  • Sandbox per run
  • No cross-tenant state
  • Structured cleanup
Calibration

Confidence is calibrated, or it's ignored

Every proposal carries a confidence score. That score is calibrated against the domain pack's regression corpus, with a declared maximum age. An uncalibrated or stale-calibrated confidence is ineligible for auto-approval — the proposal falls through to external review.

Calibration is not a per-run decoration. It is a pack-level requirement that gates the entire auto-approval path.

REGRESSION CORPUS held-out items CALIBRATE age · 12 / 30 days valid PROPOSAL 0.87 calibrated results feed back into corpus age > max → no auto-approval
Model identity

Pinned, resolved, or unknown

Every proposal records the model identity used to produce it, with an explicit version kind: PINNED (specific version locked), RESOLVED (pointer resolved at runtime), or UNKNOWN.

Only PINNED qualifies for auto-approval. RESOLVED and UNKNOWN fall through to external review, regardless of confidence, regardless of validation outcome.

PROPOSAL + model identity version kind? PINNED exact version locked → auto-approval RESOLVED resolved at runtime → external review UNKNOWN identity unverified → external review
Boundary

External authority decides

Ojas governs the run. External platform layers decide what happens next. Ojas emits typed proposals across the boundary, the external authority reads them, applies policy, and emits an approval decision. The decision is an audit artifact.

In the current internal deployment, the external authority is Aegis. Future standalone deployments will connect to a different authority that implements the same boundary contracts. The review states used are external-neutral: PENDING_EXTERNAL_REVIEW, EXTERNALLY_APPROVED, EXTERNALLY_REJECTED, NEEDS_MORE_EVIDENCE, ESCALATED, EXPIRED.

PENDING_EXTERNAL _REVIEW APPROVED terminal REJECTED terminal NEEDS_MORE loops back ↑ ESCALATED human EXPIRED timeout · no decision external-neutral review lifecycle