The standard is accepted at v0.7 on the owner's decision. §14 keeps two things
apart, as ops-warden asked: boundary assent, given by four repositories at the
version named in each record and undisturbed since; and revision review, where
all four reviewed v0.6 and every change in v0.7 is the adopted remedy of a
finding they raised. What is not claimed: nobody has reviewed v0.7 as text.
Accepting a standard nobody has re-read is deliberate. The estate will learn
more from using it than from another round of prose, and the v0.7 changes were
requested rather than invented. Findings against the accepted text stay
welcome — that is §12's normal business, not an exception.
The companion moves from canon/standards to the repository root as
SECURITY-COMPANION.md and becomes v0.2, so onboarding starts at the front door
rather than three directories down. One copy, not two: a second copy of a fact
is how the estate gets two sources for it.
v0.2 closes the gap access-engine found in v0.1 — it said publish your stance
map without saying where, and omitted the inventory obligation, so a repository
could satisfy it faithfully and no register would learn of its stance. It also
carries what v0.7 added: the corrected PEP obligations, the evidence threat
decomposition with its stated residual, cadence as MUST for load-bearing
sources with heartbeat for rare ones, the four agent rules and the glas-harness
seam, and the Railiance axes with their unsettled mapping.
It points readers at ops-warden for how to get things done. The companion says
what the rules are; ops-warden stewards the paths through them.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
Four reviews. kings-guard found that v0.6 claimed a human/agent principal
separation in §1 and §15 while §3.4 was byte-identical to v0.5: a silent edit
failure, and their framing is the right one — a rule stated about a standard in
its own change log is not a rule, which is §11's principle turned on the
standard. The same failure had dropped two §16 entries. Both restored.
§3.4 is now written: no standing credential, conduit or engine API only, agent
memory is not a state plane, every action reconstructable as the caller's, and
the seam with glas-harness for session semantics.
The rest are collisions between rules each written for its own clean case:
- §6.4(1) forbade what §6.4(3) blesses. A PEP may proceed under its declared
§9.3 stance where the application of that stance is recorded in place of the
decision — stricter than v0.6, since a fail-open result becomes metadata
rather than silence. Raised by ops-warden, the reference shape it made a
violation.
- §6.4(2) forbade the session-bound allow §9.7.1 permits. Now scoped to replay
outside the decision's own binding and lifetime, with access-engine's
canonical request digest as the mechanical test. Negative caching ruled
permitted where the refusal is recorded and the lifetime declared.
- §13.1 now exists: v0.6 mandated a stance-map register and implemented none.
Its first inventory has one row, which is the finding.
- §9.6 gained a threat decomposition after audit-core corrected its own remedy:
atomicity prevents accidental omission, cadence and reconciliation detect the
adversarial case, nothing prevents it at a compromised source. Cadence is now
MUST for load-bearing sources, with heartbeat or reconciliation required for
low-volume classes where rate monitoring cannot work.
- §9.7.2 splits by role: per input class at a PDP, one boundary deadline at a
PEP.
- §17 moves the decision-record schema to access-engine, which argued it against
its own interest.
- §13 stops attributing the actuation gap to kings-guard; §19 removed, since a
verdict inside a standard grades the document it lives in.
§20 records the Railiance interaction boundary on railiance-master's own
definitions: the four axes, that a workload is a managed running deployable so
approvals are never one, and that rein-* is not a fifth axis but a glas-harness
concern. What is unsettled is listed as unsettled.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
v0.6 is dense, self-referential, and written for readers who already live in the
estate. That is acceptable in canon and poor as a contract for the nine
repositories yet to declare and for the Staff agents expected to conform.
The companion is the operative form: the rules, no change log, no review
archaeology. Machine-readable layer key, the three sanctioned Tooling shapes as
a table, the PEP obligations, the evidence bound with the two unsound claims
written out, the four agent rules, and the four conformance states. The statute
governs on disagreement, and a disagreement is reportable as a finding.
Its §9 records that operations is HelixForge's responsibility — reef, rail,
rapp, rein — consuming the NetKingdom security and approval framework, and that
the interface is NOT yet specified. What holds today is only what holds for any
consumer. No mapping of those concepts onto the layer model should be assumed
until it is written; the statute's §16 carries the same open question.
Its §10 states the two things the estate cannot do yet — nothing is observed in
production, nothing can be contained automatically — so no reader plans around
a capability that does not exist.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
From the independent assessment of 2026-08-29, which found the model sound as a
layering constitution and incomplete as a self-healing one: cognition,
authority, and execution are specified, but the two verbs that close a healing
loop — observe in production, actuate through a deterministic surface — are
pending, and one is unstaffed.
Section numbers below §14 are unchanged; the estate cites them.
- §3.3 types the Engine layer: PDP, PIP, Evidence, Lifecycle, with a role column
in §4. Collapsing them hid different failure modes — a PIP outage is input
degradation, a PDP outage is consumer residue, an evidence-plane outage must
not block the operation it records. A new engine is a PIP unless amended.
- §6.4 names the enforcement point. The standard was precise about the decision
and silent about the gate, so enforcement lived in Staff runbooks. Four
obligations: no side effect without a decision record, no local recaching of
the verdict, a declared unreachable-engine stance, reconstructability.
- §9.2 replaced. Containment was marked pending against kings-guard, the right
mark on the wrong repository: reduce, step-up, and isolate are
authority-changing operations, so they are rendered by an Engine and enforced
by a PEP. Actuation is an unowned Engine concept held at zero. Staff proposes
containment and never performs it.
- §3.4 separates human and agent principals inside Staff — same permissions,
different blast radius. No standing credential, conduit or engine API only,
agent memory is not a state plane, every action reconstructable as the
caller's.
- §9.7 puts time into the model: explicit lifetimes, revocation visibility
deadlines, consumption as a state change never inferred from a decision
record, and the three race modes named. §9.8 states what holds under
partition.
- §17 requires the Taxonomy artifacts — claim, decision-record, gap-record, and
emission-cadence schemas — without which §6.2 and §11 are reviewable but not
compileable. Ownership proposed, not assigned.
- §18 composes the sibling standards, which had been cited in frontmatter and
nowhere in the rules.
- §5 sunsets the uncatalogued-infrastructure carve-out. §5.3 declines a proposed
fourth "operator of third-party Tooling" shape: it would convert a tracked gap
into a permanent allowance, which is the relabelling failure this standard
exists to prevent.
- §10 gains the six artifacts a layer change must carry, written from the
zone-engine case, including a permission freeze during the cut.
- §2 lifts the observation rule so it cannot be lost in a summary. §13 separates
its three normative rules from the table, now a snapshot due to move into
maturity-engine. §16 decides the approval custody question: no. §19 records
the fitness verdict.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
All four reviewing repositories returned findings on v0.4 and one contested a
rule. Every change below came from a reviewer, not from gate-house.
- §9.1 split into `pending` (no route, capability zero) and `declared-gap`
(route exists under §5.3, capability works). v0.4's single mark would have
forced a false "pending" onto ops-warden's production SSH issuance —
the fix was worse than the defect, and the defect was in this section.
- §9.3 rewritten. flex-auth contested it and was right: it collapsed "engine
reachable but degraded" with "engine unreachable", and the second has no
evaluator in the path to express anything. Input degradation is the engine's;
unreachability is the consumer's, bounded by a declared auditable total
stance — which ops-warden ADR-0009 already satisfies. v0.4 had ruled against
shipped behaviour in a repository that assented to it.
- §5 scoped: "Tooling-layer system" means a §4 Tooling row. Without this every
Staff repository was in undeclared violation for writing progress events.
- §9.4 requires the outbox to be local — no synchronous audit-core dependency
inside the state-change transaction, so an audit outage cannot block a
revocation.
- §9.5 forbids compiling maturity levels into registry content while decision
provenance carries no registry-snapshot digest.
- §9.6 gained load-bearing versus attributive evidence, the mirror rule that
absence is not evidence of non-occurrence, and kings-guard's finding that
suppression biases posture optimistic and silently.
- §11 gained a fourth state: blocked-clean, which MUST NOT rank below
conforming. A repository that declined a break-glass path and left a
capability at zero complied at cost; one that quietly opened a client and
declared nothing did not.
- §11 gained a machine-readable declaration form; ops-warden's layer.yaml is
the reference implementation.
- §13 gained state and owner-status columns; access-engine's decline of
authentication evidence is recorded.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
audit-core assented to the approval evidence half with conditions and
corrected the rationale twice. Both corrections were against wording this
standard had taken from that repository's INTENT rather than its contract.
- §4 catalogues audit-core as an Engine, on its own declaration. v0.3 named it
as an owner in §9.4 and §13 without listing it — a §11 defect in the standard
itself, which audit-core raised.
- §9.4 rationale rewritten to cite docs/integrity.md rather than INTENT
principle 6: an in-database chain detects a rewritten payload only if the
attacker does not recompute the suffix, which a database owner can, and even
with external attestation the store is not WORM or object lock. tamper
evidence is conditional on live preconditions.
- §9.4 gained emission atomicity as approval-engine's obligation, and the
prohibition on audit-core exposing an approval-validity query — a boundary
audit-core stated unprompted, applying §6.1 to itself.
- §9.6 added, estate-wide: evidence proves alteration and truncation, not
omission at source. A suppressed revocation leaves the chain intact and
verification reports intact. "The audit record proves it happened" is unsound
and is replaced with the sound form.
- §13 gained two gaps: stronger approval custody (unassigned — deciding whether
approvals need archival custody distinct from other sources is doctrine work
not yet done) and emission atomicity (approval-engine).
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
v0.2 recorded the approval object as the one open architectural hole, and
carried a latent instance of its own §9.1 rule: gate-house was catalogued as
owning conformance review with no engine to act through. Two engines were
seeded to close both.
- §9.4: the approval object goes to approval-engine — not Staff (§3.4 forbids
the runtime state), not access-engine (an evaluator owning what it evaluates
is self-dealing), not audit-core (append-only is the opposite property).
access-engine consumes approvals as input claims under §6.2; audit-core takes
the tamper-evident evidence. Operative state and evidence record are separate
artifacts with separate owners.
- §9.5: graded progression goes to maturity-engine. gate-house judges and
proposes; maturity-engine computes and remembers. Carries the guardrail that
a level may never gate a decision directly — under §6.1 that would be a
second decision point by the graded back door.
- §4 catalog gained both engines; §13 register updated.
Status proposed: the new engines are seeded by owner direction with no other
side to assent yet, and the approval evidence half needs audit-core's assent.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
All three repositories whose boundaries moved assented, each with a decision
record (flex-auth FLEX-DEC-2026-001, kings-guard KG-DEC-2026-001, ops-warden
ADR-0010), and each returned a finding. v0.2 carries the results and is
accepted; v0.1 is marked superseded and retained because the twelve estate
INTENT review notes cite it.
- §5 restructured into three sanctioned shapes: read-only diagnostics, conduit
(ops-warden's question, ruled), and declared engine gap (ops-warden's
amendment, accepted). v0.1 offered only the first, which is narrower than the
estate as it stands — a rule with no lane for a real sanctioned case gets
satisfied by relabelling rather than by closing the gap.
- §6.2 added: doctrine must reach the decision as an input claim or a versioned
policy rule. This is §6.1 applied to gate-house on the same terms it applies
to engines, drawn back by flex-auth.
- §9 added: the catalog may not assign a capability the rules forbid
discharging. Containment marked pending an engine surface; degraded-mode
fallback ruled into access-engine rather than Staff.
- §11: conformance now has three states, distinguishing a tracked gap from an
undeclared violation.
- §12 made normative, stating that an unsatisfiability finding is a success of
the conformance loop.
- §13 added: open gaps register, including the unowned approval storage and
lifecycle capability — recorded, deliberately not assigned.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
States how the security estate is layered — Taxonomy, Tooling, Engines,
Staff — distinguished by determinism and by the artifact each layer
produces, and what each layer may own.
Carries two normative rules. §5: Staff never touches Tooling directly; it
acts only through Engine APIs — the architectural form of "no privilege
from cognition", and mechanically checkable. §6: access-engine is the only
policy decision point, generalizing to the whole estate the ruling first
drawn in zone-engine/INTENT.md §5, and barring any Staff repository from
hosting a decision point.
Also fixes the vocabulary the estate has used for more than one thing:
access lane vs access rule, doctrine vs runbook, control plane as Engine
vocabulary, and the posture asymmetry.
Owner gate-house, published by net-kingdom. Status proposed: the two
adaptations carrying the most weight — flex-auth's reframing and rename to
access-engine, and kings-guard and ops-warden releasing vocabulary — are
not yet assented by their owners.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2564823@bnt-lap001
Assistant-Session: 2a7ed827-4928-4b9f-8613-9135c9cadfe9
- Fill .claude/rules/stack-and-commands.md (was an empty TODO template)
- Normalize workplan frontmatter statuses to canonical vocabulary
(completed/done -> finished) per ADR-001
- Repair glued frontmatter delimiter in NK-WP-0001 (superseded_by line)
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Initialises the net-kingdom project structure:
- README.md: updated title and description
- CLAUDE.md: project instructions and State Hub integration config
- wiki/: three reference docs (NetKingdom overview, ChatGPT and Grok
protoplans for the SSO/MFA platform)
- workplans/NK-WP-0001-sso-mfa-platform.md: combined workplan (8 phases,
8 tasks) synthesised from the two protoplans; registered in the
Custodian State Hub (workstream 39263c4b)
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
