Soul's pass on the B1 fix batch found the lift's reopen of a terminal pre-B1
continuity abort could stop Idunn booting (the reopened record claims target
authority beside a live transaction, and the store was already migrated to v4),
and that one wedged abort stopped the whole scheduler.
- The lift no longer reopens anything. A terminal pre-B1 continuity abort that
issued an activation keeps its phase and completion and carries the typed
marker CleanupEvidence::LegacySkippedBeforeB1 on its topology_reconciliation.
Validation accepts that shape only for a terminal continuity abort. Boot
reconciliation, already the exact-activation owner of that residue, is the
only cleanup path. A pre-B1 abort still in flight lifts to Pending: it
already holds its target's authority, so the lift adds no claimant.
- A failed step of a pre-fence abort is recorded as a resumable error, never
handed to the post-fence abort, which refuses a pre-fence phase and failed
every tick.
- One transaction's error no longer fails run_scheduler_tick: each candidate
is resumed on its own and its fault logged, and retiring a terminal record
tries the next one when an archive fails. An unreadable control store still
fails the tick.
- serve is boot plus the loop, so a test can run the boot sequence.
Tests pin, through the Engine: the marker and no-reopen for each pre-B1
fixture; a resident terminal abort plus a live transaction plus a second
resident abort for one target; a withdrawn incarnation never re-published and
history keeping one copy after a crash between the archive steps; a wedged
abort leaving freezing alive; an unarchivable terminal leaving resume alive;
and boot demoting the residue.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
cargo-mutants over the batch's diff missed the lift's gate, the unexplained
projection accounting, the refusal count and the deferral clock. Tests now
cover each: the lift owes only a continuity that issued an activation; a live
transaction accounts for its own activation and nobody else's; three refusals
in history hold the restart ceiling; the deferral lands between now and now
plus its length.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Soul found B1's rule sound and its transition unsafe. This makes the rule's
old inputs impossible to reintroduce a stuck state from.
- The control store's legacy lift owns the transition. A pre-B1 continuity
abort that issued an activation lifts with topology_reconciliation Pending,
so the one resolution demotes its own activation. A record already terminal
under the old rule is reopened at Starting, since a terminal record cannot
owe work. History lifts records as written. Before this, one resident
record of that shape failed validation and ControlSnapshot::read refused the
whole store, so Idunn could not boot.
- Supervision owns the Expected-only precondition. When its pre-restart
demotion fails it mints no continuity and writes next_restart_at_unix_millis
on the generation's ContinuityBackoff (attempts and window untouched), so it
neither restarts over an undemotable projection nor retries every tick.
- Boot reconciles once: a projected activation that exactly matches the
activation a failed continuity in history (or a resident terminal record)
issued is demoted. Nothing is adopted; an activation no admitted generation,
live transaction or failed issuer accounts for is reported and left alone.
New read-only CultCacheTopologyDriver::projected_activation.
- Unreadable history.cc fails closed. The continuity restart ceiling and the
consumed-command check no longer read an empty history; they stop, and the
fault is reported once while it lasts (ReportOnce). The undecodable-record
report is a pure function, pinned.
- docs/deployment-authority.md no longer says the projection never drifts.
Tests: legacy fixtures written by the 9001b58 encoder for the three abort
shapes (a second throwaway generator, see the fixtures README), driven through
the Engine; the seeded-transaction helper now takes a command kind and an
incumbent, and a Deploy carries a frozen-source receipt and a signed brake
release, so Deploy commit, Deploy pre- and post-fence aborts and a continuity
abort over an incumbent run through the real phase machine.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
The generation moves to v3 and the transaction to v4. They carry
ReadinessEvidence (the class is derived from Expected by ReadinessClass::of),
phase deadlines owned by one writer (enter_phase), RouteSupervisionState,
ContinuityBackoff, LeaseAdoptionEvidence and TerminalRecovery. The plan
moves to v3 with frozen deadline durations. A boot migration lifts every
legacy record with stored_at preserved. No decision reads the new fields
yet, so behaviour does not change at this cut.
Soul closed it after one fix batch. The first pass found every commit and
post-fence abort wedged by a deadline nothing cleared; Engine tests now
drive admissions and aborts to Complete. 204 tests. Filed forward: B3 must
give Odin its own readiness class before any route-proof admission ships,
or boot wedges; an unreadable history.cc must fail closed (B1 fix batch).
Back up control.cc before deploying: an older binary refuses a v4 store.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
advance_pre_fencing_abort survived cargo-mutants as Ok(()): the earlier
pre-fence abort test ran with candidate cleanup already complete. The seeded
transaction now takes its phase, so a Warming transaction is aborted and
driven to Complete through the stub workload port, and the abort's terminal
write is checked to carry no deadline.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
History: read_history_transactions dropped undecodable records with .ok(),
which undercounts continuity restarts for backoff. decode_history_transactions
now returns each failure with its key and reason, and the reader prints every
one plus a count. An unreadable history.cc is reported instead of read as
empty. The pre-c848459 IsolationEvidence shape needs no Legacy arm: that
commit made the enum untagged with a Linux variant of the same six fields, and
a test now decodes the old shape under both MessagePack encodings to prove it.
The Legacy structs' stated death condition now says history.cc is never
migrated or pruned, so they die only when history is.
Odin authority: current_odin_authority fell back to the bootstrap key when the
admitted Odin generation carried none (route-proof readiness). It now refuses.
OdinReceipts.authority, read only by tests, is deleted.
Tests through the production path for the rules cargo-mutants showed
unexercised: ControlSnapshot::read of a lifted legacy generation; the
route-proof arm of AdmittedGeneration::validate and route-proof shape checks;
Odin readiness, lease adoption, route supervision and continuity backoff
validation; the admitted half of max_odin_sequence; and the receipt-equality
comparisons in AwaitingReady and Routing, driven through the Engine.
The awaiting-ready default gets a comment that it is unsafe for stateful
targets whose migrations run in AwaitingReady; B5 owns the fix.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Every admission commit and every post-fencing abort failed validation:
transition() stamped a phase deadline on entry to Fencing..Committing and
nothing cleared it, so the record cloned into Complete still carried a
Committing deadline and validate() refused it. The first deploy or continuity
past Fencing wedged in Committing; a post-fence abort never completed.
DeploymentTransaction::enter_phase is now the only writer of phase. It sets
the phase, the update time and the deadline together, so a record outside
Fencing..=Committing carries no deadline by construction. transition(), both
abort completions, the commit, rejected() and Starting/Warming all use it.
Wall-clock ordering rules are deleted: entered_at <= updated_at, created_at <=
updated_at and promoted_at <= updated_at each wedged a transaction after an
NTP step backwards, since same-phase writes stamp now_millis(). The deadline
is entered_at + duration, fixed at entry.
The Engine takes its systemd workload port as a trait object
(Engine::open_with_systemd_workload), which lets three Engine-level tests drive
a continuity transaction from Fencing through Leasing, AwaitingReady, Routing
and Committing to Complete, and a post-fence abort to Complete, with real
Odin-signed evidence and no systemd, nginx or network. deployment_plan's test
fixtures become pub(crate) for them.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Soul accepted it on 2026-09-29. With an exact fragment, a stale
observation costs one challenge and zero actuations; rendering is
byte-deterministic; drift restores once. Its merge with idunn/cut1-fix5
is clean (213 lib tests). The storm paths it does not claim go to B2.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The fixture envelope's stored_at now differs from the record's own update
time, so a migration that re-stamps the envelope fails the test.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
One migration for the route-continuity foundation. No decision reads the new
state.
- ReadinessEvidence (OdinCorrelated | RouteProof) replaces the bare Odin Ready
receipt on both records. The class a target must collect is derived from its
Expected (ReadinessClass::of): a shared-infrastructure odin.verse-rendezvous
dependency means Odin-correlated, none means route-proof. It is not stored.
RouteProof evidence is refused for a target that declares Odin.
- Generation carries route supervision state (challenge times, failures,
degraded marker, actuation window) and continuity backoff. A new generation
inherits the actuation window and backoff and starts unobserved and
undegraded. route_repair_started_at stays readable.
- Transaction carries the current phase deadline (stamped by transition from
the plan's frozen durations) and lease adoption evidence.
- FailedAfterFencing names its recovery: RestartAdmitted, RestoreIncumbent
(the default for older records) or OperatorRequired{reason}.
- Every v2 generation lifts as Odin-correlated with its receipts unchanged;
v2 and v3 transactions lift through LegacyDeploymentTransaction. The boot
migration rewrites both, preserving stored_at, and history reads lift the
same way. The Legacy structs die when no older record can remain.
Fixtures under tests/fixtures/idunn-control-legacy are the bytes the 9001b58
encoder wrote for records built through its own constructors.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
An operator binding may now declare a [deadlines] table naming any of the
five post-fencing phases (fencing, leasing, awaiting_ready, routing,
committing) in seconds, 1..=86400. compile_deployment_plan resolves the
binding over Idunn's defaults and freezes the result in the plan, so a later
change to a default cannot move an in-flight plan.
The plan becomes schema v3, with the frozen durations as a trailing field.
A v2 plan keeps its nine-field encoding (the field is skipped when absent),
so every existing plan id still verifies; it runs on Idunn's defaults at read
time. Nothing reads the durations yet.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
A continuity candidate shares the admitted Expected key. Both abort intents
marked its topology reconciliation Skipped, so a failed continuity left its
candidate activation projected under the admitted key while the admitted
receipt named the older one, and a later deploy's post-fence demotion refused
as substituted.
DeploymentTransaction::abort_topology_reconciliation is now the one rule for
both intents and pre-fence validation: a deployment owes the projection when it
published its Expected, a continuity owes it when it issued an activation.
reconcile_failed_candidate_projection is the one resolution for both abort
paths: a deployment withdraws its own key, a continuity demotes the shared key
to Expected-only with its own exact activation and never withdraws the
Expected. The post-fence incumbent demotion stays deployment-only and exact.
Tests read the persisted topology store after the production
publish/demote sequence.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
This reverts commit a148802cb6.
Rollback after fencing adopted whatever activation was projected under the
shared Expected key, so a continuity abort's leftover activation could be
demoted by a later deploy. The drift is closed at its source by the next
commit: a continuity abort demotes its own exact activation. Rollback demotes
the admitted generation's exact activation again.
The doc paragraph is dropped; the 9f00e7a paragraph above it stays until B4.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Delete the stale-observation -> repair-intent -> restore path from
supervise_admitted_route. An exact fragment with a stale observation now
yields one route challenge and nothing else. restore_admitted_membership
returns before nginx -t, the private-mount unit, ufw and reload when the
fragment already equals the admitted membership. route_repair_started_at
stays readable but no longer decides anything.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
All six were ruled as recommended, one at a time, on 2026-09-29. Q6 adds
a Soul pass on the weekend CultLib branch before it merges. The campaign
moved to the Eureka session "Codebase audit".
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Body facts, identity/lifecycle/authority table, six operator questions and a
draft cut list (S1, B1, F0, B2-B5, S2, C1, S3/C2) for the route-continuity
target. Campaign handed to the "Codebase audit" Self session; operator forks
are ruled there.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Eureka target for the route-continuity rebuild after the 2026-09-28 StreamPixels
deployment failure: observation never actuates routing, failure never
accelerates actuation, web apps prove over their own HTTP route, every
post-fencing phase ends.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The first v2 Ghostlight deploy sealed its release and then failed creating
the runtime bundle: /etc/gamecult/ghostlight/runtime was read-only under
ProtectSystem=full. Runtime root only, not the credential directory.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The version tag could not be recorded where the ruling said, because the
stored value is validated to a bare sha256 shape by a pinned dependency in
another repository. Hands implemented an equivalent implicit version read,
flagged the divergence, and asked whether proceeding was right. It was: the
ruling named a mechanism, the hazard was live, and halting would have left the
fleet exposed.
Five tests prove it by failing without the fallback, which is the point, since
the whole finding was that nothing had ever exercised the stored path.
The frozen-source digest turned out to have the same exposure through a
narrower window, and this cut's own sixth and seventh batches crossed it
twice. Same answer applied, with the prior shapes reproduced from git history
rather than from memory.
Hands also spotted a third instance and declined to chase it: the systemd
actuator compares a raw content-only digest read through procfs against a
stored value now carrying mode, gid and xattr terms. Two different algorithms
compared to each other, which is the same defect a third time. Flagged for its
own look rather than guessed at.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The bypass is a compile error now, proven by builds that fail on Soul's exact
spellings rather than by reading the lint config. The repository had no CI
surface at all before this, so there was nothing for "denied in CI" to attach
to until now.
Hands flagged that the single-file artifact digest format changed and reasoned
nothing compares against literals. That holds for fixtures and not for
receipts: two call sites in host.rs compare a freshly computed digest against
a stored value. So the first upgrade past this commit makes every already
installed workload fail "installed executable differs from Expected", and
every running one fail "the installed executable changed under the running
process". Idunn is the deployment and daemon-survival authority, and this
would land silently at restart across the fleet.
The ruling versions the digest and verifies in the version the receipt names,
rather than re-sealing admitted state to make new code agree with itself. It
is not closed until a test seals under the old format and verifies under the
new, because the whole finding is that nothing exercised the stored path.
Two bugs in this batch were found by running the suite rather than trusting
it, including a test that set an attribute on a root whose own metadata is
never hashed.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
The port can be walked around. Two spellings of the original bug restore it
with all 193 tests green, and the proof it leaks is already in the file: one
function still calls fs::read_link directly, and a probe measured sixteen
counted calls beside three uncounted syscalls. My record claimed the count was
total rather than conventional. It is still conventional; the convention moved
from remembering the counter to remembering the port. That is the second time
this record has overclaimed what a pin does.
The artifact digest's single-file path never got the ruling, and it is the
path the installed executable actually takes. Setuid, mode and a file
capability are all invisible there, immediately before launch. Every test for
it wraps the file in a directory so the other path runs, and one of their doc
comments claims to pin exactly the scenario it does not reach.
The xattr framing is sound and nothing proves it. The suite stays green with
values dropped entirely, so a capability edited from cap_setuid to
cap_sys_admin is covered by no test.
Soul named one cause with three faces: every pin here is a convention a later
edit steps around, and the fixtures use the convenient shape rather than the
one production uses. So the eighth batch makes the bypass a compile error and
moves the tests onto the production shape, instead of adding a ninth round of
point fixes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
193 tests, and the mutation that passed in 0.39 seconds on pass 7 now fails.
Putting canonicalize and symlink_metadata behind a port that charges the
counter inside the same method as the syscall is what turned an honour system
into a measurement; the hand-placed call sites are gone.
The validator mask tests are precise rather than merely red: narrowing the
file mask fails exactly the setgid test and narrowing the directory mask fails
exactly the setuid one, with the other five staying green. That is the
evidence they pin single bits.
The unreachable test is deleted rather than repaired, with a note saying why
the defence it claimed never existed. The xattr tests close Soul's file
capability finding and state in their own doc comments that the probe runs as
root and proves nothing about an unprivileged attacker.
Hands shipped a compile bug in the first commit and its own Yggdrasil run
caught it before the report. That is the order these should happen in.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Soul restored the exact bug this line of rulings exists to prevent, without
instrumenting it, and the new call-count pin passed in 0.39 s. The counter
only counts calls someone remembered to annotate, so it is an honour system
rather than a measurement. The code said so in a doc comment four lines below
the assertion; my record claimed the opposite. That record is corrected here.
The fixture I recorded as deleted for being unable to fail is in the tree,
still unable to fail. Hands' reasoning about why the collision class is
unreachable was sound and Soul retested it; what landed on top of that
reasoning was a test that cannot go red.
A file capability is invisible to both the validator and the digest, which is
the grant setuid used to be with no mode bit to see it. no-new-privileges
prevents it on our own deployment path, but the digest is the detection and
Cut 4's verdict claims the tree is the one we froze. The artifact digest also
got a single executable boolean where the ruling said full mode and gid, so a
post-install setuid re-verifies clean.
Six rulings for the seventh batch. The first moves filesystem access behind a
narrow injected port so the count is total rather than conventional, which is
the shape doctrine already asks for.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
183 tests green. The timing pin is now a call counter asserting an exact
formula, which is what it should have been before a clamp made a 1.63x
separation look like 33x.
I ruled the bind mount should carry nosuid. Docker exposes no such
per-bind-mount flag through either syntax, and Hands verified that against
the real CLI rather than arguing from memory. The container already runs with
no-new-privileges, which is process-wide and stronger than what I asked for.
The ruling is withdrawn: the prevention was already there and I ordered a
second one without looking.
Two fixtures in this batch would have lied and were caught by mutating rather
than reading. The container's chown clears a freshly-set setuid bit, so a
fixture that sets then chowns tests nothing. And the first observe_frozen test
tampered with the recipe file, which a different check inside observe_frozen
catches, so it stayed green with the digest check deleted.
The artifact digest's symlink terminator is deliberately left unpinned:
symlink(2) cannot put a NUL in a target, so no honest test exists. Hands built
one, confirmed it did not go red, and deleted it instead of keeping it.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Five of seven rulings hold. The containment argument, where I asked Soul to be
most suspicious after S5-8, survived twenty escape shapes it built itself.
R-I1's pin is not a ratio test. Its clamp fires on every run, so it reads as
"under 200 ms", and clean code has already measured 198.9 ms under the
three-slot load the operator just authorised. The reported 33x separation was
an artifact of that clamp; the real separation is 1.63x, which a slightly
faster host erases. It fails in both directions, and under Cut 4 a false red
is a false verify verdict on unattended work.
R-I3 is not met. The validator masks the mode to 0o777 and checks uid but not
gid, so setuid, setgid, sticky and gid are all permitted to vary and none are
hashed. A frozen binary made setuid-root re-observes clean against its
receipt, and the tree is bind-mounted into runner containers with no nosuid.
The ruling adds the mount flag as well as the digest terms: the digest detects
and the flag prevents, and the flag costs a mount option.
Soul also found that the artifact digest still has the blindness the frozen
source digest just lost, and that observe_frozen, which is the tamper check
itself, has no test at all.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
All seven rulings landed, 175 tests green, and every new test dies under its
own mutation. The two-link cycle fixture hangs rather than asserting, which is
the correct falsification: nothing but the budget bounds a cycle, so a timeout
is what a broken budget produces.
The batch's real finding runs the other way. Soul called three containment
checks redundant and I ruled delete them. Hands deleted two, hand-traced the
third, and found that with all three gone a symlink whose whole target is ".."
at the root resolves to the parent of the root and returns Ok, because no
later Normal step ever runs to trip the other arm. That is a containment
escape, the exact class this resolver exists to prevent. Soul was wrong, my
ruling repeated it, and a hand-trace caught it.
Soul's HTTPS rig had also evaporated: no session, no work directory, no
scratch trace. Hands rebuilt it from this map's description, which worked only
because the description was detailed. A probe that is going to be committed
has to be handed over while it still exists.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Soul could not falsify any claim the fourth batch made. What it found instead
is that the S4-1 fix reached the exponent and not the per-component cost, and
that the line responsible is dead code: the value is already canonical, so
deleting four lines removes a cubic term and leaves the suite green.
The other half of the bill is coverage. The traversal budget is the only thing
that terminates a symlink cycle and nothing tests it; cargo-mutants missed two
mutants on that single line. The digest that serves as the tamper check is
blind to symlink targets and to the executable bit, both of which the
validator permits to vary. S4-6 was reported landed and has no test at all.
Seven rulings for the fifth batch. R-I5 departs narrowly from my earlier
"keep S6's behaviour exactly" and says so rather than letting it pass as
drift: the 40-traversal ceiling matches the kernel's own ELOOP limit and is
kept, but it must be declared and the error must name the link.
FU-Heimdall-Gitlink cleared upstream on its own. Heimdall froze end to end,
945 entries, zero mismatches. The ops follow-up closes.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Soul's fourth pass found the symlink resolver newly exponential
(resolve_frozen_source_symlink recursed per path component and
re-resolved every shared prefix from scratch, so an n-link
self-doubling chain cost 2^n real filesystem resolutions), the
streamed hash unpinned, three of the five S5 guard tests not killing
their own guard, all three S1 fsck guards unpinned against
already-local objects, no test proving a legitimate in-root ".."
freezes, and the symlink writer still buffering a whole blob for a
120000 entry.
S4-1: resolve_frozen_source_symlink is now an explicit iterative walk
(a work-stack of frames standing in for the old call stack) with
per-symlink memoisation and one shared traversal budget charged per
distinct symlink read, not per nesting depth. Soul's n=20 doubling
chain (a0 -> ., ak -> a{k-1}/a{k-1}) now resolves in milliseconds
instead of 43.7 s, and freezes rather than refusing: nothing in the
chain dangles or escapes the root, matching S6.
S4-2: two trees differing only past the first 64 KiB chunk now assert
different digests, killing the zero-length-buffer and first-chunk-only
mutants cargo-mutants found.
S4-3: the duplicate-path, ensure_frozen_directory containment, and
create_dir guards each get a fixture that only that guard refuses --
calling git_tree_entries directly (before any fsck runs) for the
duplicate path, and pre-seeding created_dirs with state a second,
independent bookkeeping pass could produce for the other two.
S4-4: Soul's eleven already-local hostile fixtures (dup-trees, six
.git look-alikes, an unsorted tree, mode 120000 on a tree, mode 040000
on a blob, a zero-padded mode) land as tests against freeze_exact's
own explicit fsck. A second test proves freeze_exact's own fetch fsck
flag is what refuses a hostile revision that was never local, by
asserting which context wraps the error.
resolve()'s own fetch fsck flag is still unpinned: exercising it needs
a real HTTPS origin (OperatorBinding::validate requires one), and
url.insteadOf does not work around that -- `git remote get-url` expands
the rewrite, so it can never satisfy ensure_checkout's own
origin-equality check. Confirmed by hand against a local git; a real
smart-HTTPS harness like Soul's S1 rig is what closing this needs.
Recorded as a gap, not faked with a broken test.
S4-5: a legitimate in-root ".." link (two hops up from a nested
directory, back to a file at the root) now has a test proving it
resolves and freezes, not just that an escaping "..' refuses.
S4-6: materialize_tree_raw refuses a 120000 entry whose blob exceeds
the platform's PATH_MAX before reading it, for both the streamed-blob
and the shares-content-with-a-regular-file cases.
Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>