Draft WARDEN-WP-0032 properly — security zones, ownership question first
All checks were successful
CI Smoke / host-smoke (push) Successful in 0s
CI Smoke / container-smoke (push) Successful in 1s

Expands the placeholder into a real plan, grounded in what the estate already
has rather than first principles:

- tenancy-posture_v0.1 is the structural model to follow (canon standard +
  per-repo declaration + "accuracy, not altitude"), and its §14 adoption stance
  is "structure, not tooling" — a precedent about sequencing.
- Reefs are substrate placement zones with real risk statements. A security
  zone is not a reef; repo-manager's own "topology is not readiness" applies
  here as "placement is not posture".
- flex-auth is latency-critical and already consumes a compiled registry
  snapshot, so zone membership should reach the PDP by compilation rather than
  a synchronous lookup in the decision path.
- warden plan already returns verdicts with reasons; zone-aware enforcement
  extends that machinery instead of growing a parallel one.

T01 settles ownership before any modelling, because it decides which repo the
rest lives in: canon standard, a zone-engine runtime authority, or flex-auth
policy data. The recommendation is canon-first with zone-engine named as an
explicit promotion — but the time-boxed refactor exception is the case that
could justify an engine immediately, since an expiring grant is state and state
wants an owner.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
This commit is contained in:
tegwick 2026-08-19 20:38:25 +02:00
parent 3d968bb993
commit b845f4b51e
2 changed files with 180 additions and 63 deletions

View file

@ -1,7 +1,7 @@
---
id: WARDEN-WP-0032
type: workplan
title: "Security zones — scope enforcement to bands of differing rigidity"
title: "Security zones — bands of enforcement rigidity, and where they live"
domain: infotech
repo: ops-warden
status: proposed
@ -18,48 +18,103 @@ updated: "2026-08-19"
# WARDEN-WP-0032 — Security zones
`ADR-0006` defers `policy.enabled: true` until enforcement can be scoped. This
workplan builds the thing it defers to.
`ADR-0006` defers `policy.enabled: true` until enforcement can be scoped to a
zone. This workplan builds the model it defers to, and — first — settles who
owns it.
## The problem in one paragraph
## The problem, precisely
ops-warden has three posture axes already: environment (`dev`/`test`/`prod`,
WP-0015), workload maturity (`M0``M3`, WP-0015), and organization posture
(`build`, WP-0029). What it does not have is a way to say *"this band of the
estate is enforced, that one is advisory"* — so every enforcement control it
ships is a repo-wide boolean. `policy.enabled` is the first one to actually
become flippable, and flipping it would harden the access needed to perform the
refactors currently underway. The axes describe posture; nothing acts on them.
ops-warden can express *how exposed* something is and *how mature* it is. It
cannot express *how rigidly this band of the estate is governed right now*.
## What a zone has to answer
| Axis | Shipped | Answers |
| --- | --- | --- |
| Environment posture (`dev`/`test`/`prod`) | WP-0015 | Where does this run? |
| Workload maturity (`M0``M3`) | WP-0015 | How ready is this workload? |
| `organization_posture: build` | WP-0029 | What state is the org in? |
| **Enforcement stance** | **missing** | **Is this control enforced here, and what happens when it fails?** |
A zone is only useful if it decides something. Minimum:
Because the fourth is missing, every enforcement control ops-warden ships is a
repo-wide boolean. `policy.enabled` is the first to become flippable, and
flipping it would make flex-auth a hard dependency of every `warden sign`
including the certificates the ops-bridge tunnels depend on, one of which
carries the policy call itself. Across an estate mid-refactor that hardens
exactly the access needed to perform the refactor.
- **Membership** — what puts a lane, actor, or workload in this zone? Derived
from the existing axes where possible rather than a fourth hand-maintained
list.
- **Enforcement stance per control** — for each gate (pre-sign policy, agent
read-boundary, high-risk refusal): enforced, advisory, or exempt.
- **Failure mode** — fail-closed or fail-open, per control per zone. This is the
actual bite: `fail_closed` is what makes a dead tunnel an outage.
- **Who may change it, and how it is observed** — a zone that can be quietly
widened is not a boundary.
The three existing axes *describe*. Nothing *decides*. A zone is the object that
decides.
## Open questions to settle before designing
## What already exists — do not reinvent these
- Is a zone a property of the **actor** (`adm`/`agt`/`atm`), the **lane**
(catalog entry), the **workload**, or the **moment** (an active refactor)? The
CoulombCore decommission suggests the last one matters and is the hardest to
model — a time-boxed relaxation is a different object than a standing zone.
- Do zones compose with the three existing axes, or replace one? A fourth
independent axis multiplies states; folding `organization_posture` in may be
the honest move.
- Is this ops-warden's model to own at all, or NetKingdom canon that ops-warden
consumes? ADR-0005 says implement narrowly and route broadly. A security zone
model that only ops-warden honours is worth much less than one flex-auth,
secrets-engine, and tenant-engine also read — but ops-warden should not
absorb an estate-wide model on its own authority either. **Settle this
first**; it changes who writes the rest.
Four estate mechanisms are adjacent, and getting the relationships right is most
of the design work.
**`net-kingdom/canon/standards/tenancy-posture_v0.1.md`** — the closest
structural analogue and the model to follow: six orthogonal axes, graduated
levels, a per-repo declaration file (`tenancy.yaml`) carrying current /
implemented / target vectors with evidence and a review date, and §6
"conformance is accuracy, not altitude". Its §14 adoption stance is
*"structure, not tooling"* — a direct precedent about sequencing, and one this
workplan should not quietly break.
**Reefs** (`repo-manager/docs/RailianceAppDeploymentGuide.md`) — `reef-railiance`,
`reef-storage`, named in a rapp's `bound_reefs`. These are **substrate**
placement zones with real risk statements attached (`reef-railiance` carries a
single-node/shared-control-plane residual risk requiring explicit acceptance).
A security zone is *not* a reef — reef answers *where does it run*, zone answers
*how rigidly is it governed* — but they are correlated, and a taxonomy that
ignores reefs will fight them. Also note repo-manager's own ADR-0006: *topology
is not readiness*. The same caution applies here: **placement is not posture.**
**flex-auth** — the PDP, `service_class: latency-critical`, which already
consumes a *compiled registry snapshot*
(`registry/flex-auth/production_registry_snapshot.json`, built by
`scripts/build_flex_auth_registry.py`). This matters enormously for T03: zone
membership can reach the decision point by compilation, with no runtime lookup
and no new service in the request path.
**`warden plan`** (WP-0029) — already classifies a need as `autonomous` /
`founder_required` / `unroutable` from properties of the thing, returning the
`reasons` that produced the verdict. Zone-aware enforcement should extend this
verdict machinery rather than grow a parallel one.
## The question that changes everything else: who owns this?
Three placements are defensible. **T01 settles it before any modelling**,
because the answer changes who writes T02T05 and in which repo they live.
**(a) NetKingdom canon + per-repo declaration.** A `security-zones_v0.1.md`
standard beside `tenancy-posture_v0.1.md`; each repo declares its lanes'
zones in a `zones.yaml` the way it already declares `tenancy.yaml`; flex-auth
enforces from the compiled registry. ops-warden drafts and offers, exactly as it
did for the WP-0015 maturity model. *Strongest fit to precedent; no new runtime
dependency; matches "structure, not tooling".*
**(b) A new `zone-engine` repo.** A runtime authority owning zone identity,
membership, and lifecycle, in the family of `tenant-engine` / `user-engine` /
`secrets-engine` — each of which has a boundary contract in canon. *Justified
if and only if zone membership needs minting, an API other components query, or
— the real candidate — **time-boxed exceptions with expiry and an audit trail of
who widened what**. A standing classification does not need an engine; a
relaxation that must expire probably does.*
**(c) flex-auth policy data.** Zones as attributes in the policy package, no new
artefact anywhere. *Cheapest, and wrong if zones must govern controls flex-auth
does not evaluate — the agent read-boundary (ADR-0004) and `warden desk`
escalation are ops-warden-side and would fall outside.*
**Recommendation to be tested in T01, not assumed:** start at (a), and name
(b) as an explicit promotion with a written trigger, rather than seeding an
engine before the model exists. An empty `zone-engine` would end up defining
zones by accident — through whatever its first schema happened to be — which is
the failure `ADR-0001` names in a different register. The counter-argument
deserves a fair hearing: if time-boxed refactor exceptions are in scope from day
one, (b) may be right immediately, because an expiring grant is state and state
wants an owner.
Under `ADR-0005` ops-warden does not absorb an estate-wide model on its own
authority. A zone model only ops-warden honours is worth much less than one
flex-auth enforces and secrets-engine reads.
## Tasks
@ -69,11 +124,15 @@ status: todo
priority: high
```
**Decide ownership before designing.** Take the third open question to
net-kingdom (canon) and flex-auth (the PDP that would enforce it). If it is
canon, ops-warden drafts and offers, as it did for the WP-0015 maturity model,
and consumes the result. If it is ops-warden's, say so explicitly and record
why. Do not build a model whose owner is unsettled.
**Settle ownership and placement.** Take the three options above to
`net-kingdom` (canon) and `flex-auth` (the PDP that would enforce), with the
time-boxed-exception question stated explicitly since it is what decides
between (a) and (b). Record the outcome as a decision with its reasoning — if
it lands outside ops-warden, this workplan's remaining tasks move to the owning
repo and ops-warden keeps only its consumer-side tasks.
Deliverable: a decision record, and coordination messages sent. Do not model
anything before this closes.
```task
id: WARDEN-WP-0032-T02
@ -81,12 +140,24 @@ status: todo
priority: high
```
**Name the zones against reality, not in the abstract.** Walk the current
estate — the 27 catalog lanes, the actor inventory, the workloads with posture
descriptors — and see what bands actually exist. A zone model that does not
partition today's lanes cleanly is wrong, and the exceptions are the interesting
part. Include the refactor case explicitly: what zone is a lane in while the
component it fronts is being rebuilt?
**Derive the zones from the real estate, not from first principles.** Walk the
27 catalog lanes, the actor inventory (`adm`/`agt`/`atm`), and the workloads
carrying posture descriptors, and find the bands that actually exist. The test
of the model is that it partitions today's estate without a residue of
exceptions; the exceptions found are the most informative output.
Must answer explicitly:
- Is a zone a property of the **actor**, the **lane**, the **workload**, or the
**moment**? The CoulombCore decommission shows the last one is real and is the
hardest to model — a time-boxed relaxation is a different object from a
standing zone, and conflating them produces a permanent hole with a
temporary-sounding name.
- Do zones **compose with or fold in** the three existing axes? A fourth
independent axis multiplies states; absorbing `organization_posture` may be
the honest move.
- How do zones relate to **reefs**? Correlated, not identical. Say so in the
model rather than leaving readers to assume either.
```task
id: WARDEN-WP-0032-T03
@ -94,22 +165,35 @@ status: todo
priority: high
```
**Model enforcement stance and failure mode per zone per control,** covering at
minimum: the flex-auth pre-sign gate (`policy.enabled` / `fail_closed`), the
agent read-boundary on `risk: high` lanes (ADR-0004), and the `warden plan`
escalation verdicts (WP-0029). Each must be expressible as enforced / advisory /
exempt with an explicit failure mode.
**Model stance and failure mode per zone per control.** For each control:
enforced / advisory / exempt, *and* fail-closed or fail-open. The failure mode
is the part with teeth — `fail_closed` is what turns a dead tunnel into an
outage, and it is the reason `policy.enabled` was deferred.
Controls in scope at minimum: the flex-auth pre-sign gate (`policy.enabled`,
`fail_closed`), the agent read-boundary on `risk: high` lanes (`ADR-0004`), and
the `warden plan` escalation verdicts (WP-0029).
Prefer **compilation over runtime lookup** for anything flex-auth evaluates: it
is latency-critical and already consumes a built registry snapshot. Adding a
synchronous zone lookup to the decision path would be a regression dressed as an
architecture.
```task
id: WARDEN-WP-0032-T04
status: todo
priority: medium
priority: high
```
**Make `policy.enabled` zone-aware and retire the global boolean.** The
successor control reads the zone of the actor being signed for. Ship the
migration in the same change as the deprecation — leaving both is exactly the
second-source-of-truth failure ADR-0001 exists to prevent.
**Declaration format and conformance rule.** Follow `tenancy.yaml`: a per-repo
declaration, with evidence, a `reviewed` date and a `review_due`, and the §6
rule carried over — **accuracy, not altitude**. A repo declaring a stricter zone
than it can evidence is the failure mode to design against, because it is the
one that looks like progress.
Include how a zone assignment is *changed*, and how that change is observed. A
zone that can be quietly widened is not a boundary. If time-boxed exceptions
exist, their expiry must be enforced by something, not by an intention.
```task
id: WARDEN-WP-0032-T05
@ -117,7 +201,34 @@ status: todo
priority: medium
```
**Close WARDEN-WP-0031 T05 through this model** — enable the pre-sign gate for
whichever zone the SSH signing lane lands in, with that zone's failure mode.
Re-run `scripts/check_policy_caller_identity.py` first; the WP-0031 evidence
will be stale by then and re-establishing it is cheap by design.
**Make `policy.enabled` zone-aware and retire the global boolean.** The
successor control reads the zone of the actor being signed for and that zone's
failure mode. Ship deprecation and migration in the same change — leaving both
is the second-source-of-truth failure `ADR-0001` exists to prevent.
This closes `WARDEN-WP-0031-T05`, cancelled under `ADR-0006`. Re-run
`scripts/check_policy_caller_identity.py` first: the WP-0031 evidence
(`decision:f3f7c88f9585582a`, 2026-08-19) will be stale, and re-establishing it
is cheap precisely so this is a re-check rather than a re-do.
```task
id: WARDEN-WP-0032-T06
status: todo
priority: medium
```
**Close out honestly.** Update `SCOPE.md`, `wiki/WorkloadSecurityPosture.md`,
and `wiki/PolicyGatedSigning.md`. If the model landed in canon or in another
repo, `ADR-0006` gains a successor or an amendment recording where enforcement
scoping now lives — ops-warden follows it rather than owning it, and the ADR
must say which.
## Deliberately out of scope
- **Network segmentation.** "Zone" is an overloaded word; this model is about
*enforcement rigidity*, not routing or NetworkPolicy. Say so wherever the term
first appears, or the two will be conflated within a quarter.
- **Re-litigating the three existing axes.** They stay as they are unless T02
finds a fold-in.
- **Zones for anything ops-warden does not consume**, unless T01 hands the model
to an owner who needs the wider scope.