Verified Closure and Evidence

Activity becomes trusted work only when the result can be proved.

Coda preserves the full chain from observed condition to authorized action, resulting condition, independent verification and closure.

Closure anatomy

Seven elements make an exception inspectably complete.

01

Observed condition

What was wrong or outside the operating contract.

02

Mandate

Why the guild was authorized to intervene.

03

Responsible owner

The accountable sponsor and operating role.

04

Authorized action

The response permitted by the authority model.

05

Resulting condition

The system state after intervention.

06

Independent verification

Proof the intended outcome was actually achieved.

07

Evidence Pack

The replayable and exportable record of the complete closure.

Evidence Pack

Portable proof, not platform reassurance.

The pack should remain understandable after a model version, runtime, operating team or vendor changes.

Authority

Who acted for whom

Principal, role, delegation, scope, purpose and expiry.

authority.path = sponsor → guild → operator
Decision

Why the response was selected

Policy, risk tier, human approval and exceptions.

decision.result = DENY_AND_SUSPEND
Action

What changed

Tool invocation, resource, arguments, side effects and rollback.

action.side_effect = none
Outcome

What was independently verified

Before state, resulting condition, verifier and acceptance criteria.

closure.verified = true

Replay

Reconstruct the closure without trusting the original operator.

Replay should expose the event, mandate, evidence available at the time, authority decision, action path, resulting state and verifier acceptance.

01

Rebuild context

Load source evidence, policy versions and operating contract.

02

Re-evaluate authority

Confirm identity, delegation, scope and expiry.

03

Inspect effects

Trace every tool call and resulting system mutation.

04

Reproduce acceptance

Apply closure criteria to the resulting condition.