---
slug: policy/active
version: "0.1"
title: Policy — Active
summary: Read the policy currently active for a purpose decision slot — the active binding for one purpose, or every active binding. The read companion to policy/activate.
status: draft
targetFrameworkVersion: "0.5.0"
category: governance
keywords:
  - policy
  - rego
  - active
  - governance
authors:
  - Glenn Gore (https://github.com/stormer78)
parties:
  - role: vault consumer
    requirement: REQUIRED
    member: issuer
  - role: vault maintainer
    requirement: REQUIRED
    member: recipient
proofRequirement:
  requirement: RECOMMENDED
  rationale: Read-only inventory of the active bindings. Recommended for attribution.
sideEffects:
  level: none
  rationale: "Read-only; returns the current active bindings and persists nothing."
subjectPath: /purpose
exposure:
  discloses: metadata
  actsAsSubject: false
errorCodes: []
---

## Abstract

The **Policy — Active** Trust Task reads the active-policy bindings — the `(contextId, purpose) → policy` map that [`policy/activate`](../../activate/0.1/) writes. It is the read side of that write: given a `purpose`, it returns the single policy currently authoritative for that slot (or nothing, if none is active); with `purpose` omitted, it returns every active binding.

It exists as its own task rather than a filter on [`policy/list`](../../list/0.3/) because a purpose is **not** a property of a policy in this model — it is a binding produced by activation (see `policy/activate`). `policy/list` enumerates policy *modules* (which carry `appliesTo`/`priority`, not `purpose`); this task enumerates the *bindings*. A `purpose` filter on `list` would be answering a question its result shape cannot express.

## Conformance

Producer: optionally supply `purpose` to narrow to one slot, and `contextId` to scope to one context.

Consumer: verify `PolicyAdmin` capability (the same gate `policy/list` applies). Return one `ActiveBinding` per matching active slot, each carrying the full `PolicyModule` so the caller need not follow up with `policy/get`. When `purpose` is supplied and no policy is active for it, return an empty `bindings` array — not an error; absence of an active policy is a normal state, not a lookup failure. The set is unpaginated: there is at most one active policy per purpose, so the active set is small by construction.

## Security & Privacy

**Source disclosure.** Each binding carries the active policy's `PolicyModule`, including its Rego `module`. As with `policy/list`, Rego source is not a secret but describes the maintainer's security posture and SHOULD be visible only to admin-class consumers — the `PolicyAdmin` gate enforces that.

**No existence oracle over purposes.** An unknown or unbound `purpose` returns an empty array, identical to a known-but-inactive one, so this task does not distinguish "purpose never used" from "purpose has no active policy". That is intentional: `purpose` is a free maintainer-scoped string, and there is nothing to enumerate against.

**Auditing.** Recommended at sampled rate, matching `policy/list`.

**Free text.** `purpose` names a decision slot and is constrained by no
vocabulary the framework defines, so it is free text under
[SPEC.md §7.3](/SPEC.md#73-specification-requirements) item 19. The request
member already declared `maxLength: 128`; the `purpose` inside each returned
binding now declares the same bound. It is authored by the operator who
activated the binding, read by the maintainer's policy router and by whoever
reads this listing, and **retained** for as long as the binding is active,
because it is half the binding's key. It carries no authority: a caller decides
nothing from the slot name alone.

