Do not turn an empty cached digest into an unbounded descriptor hash. Distinguish size refusals from YARA results so the exec guard applies its own bounded scan/degrade policy. Cover cached read budgets across size boundaries, known hashes, inventory/live events and real guard-cache reuse. Fold/live/KPI and post-rollout CPU acceptance remain pending; OpenSnitch checksum work is separate.
Use suspend-aware BOOTTIME waits and owned-container cleanup without changing the workload. Reject missing coverage and preserve producer/report failures. Include runtime/helper regressions, constrained build-job support and live exec-guard crash safety. Document the failed live/idle gates without relaxing thresholds or claiming production acceptance.
Preserve independent support, typed rule namespaces and stable process lifetimes. Share authoritative managed-set reconciliation across daemon and offline correlation, load offline policy without seeding/migration, and release resources after shutdown-flush exceptions. Keep retirement history, transaction ownership and fault/allocation regressions; no whole-set crash-atomicity or live acceptance claim.
Reuse the existing resolver/sampler for opt-in aggregate training summaries. Reject incomplete or undecodable snapshots without usable partial reports, retain best-effort readers elsewhere, and document cooperative extent checks rather than atomic snapshot guarantees. Include command-table/help, JSON, fixture and corruption regressions; no model training or promotion.
Represent full-range ordered clock differences without signed overflow and allow complete decay. Add seeded restore-clock regressions and a dedicated fuzz seed.
Allowlisting a binary suppressed only *new* detections: the findings
already stored for it stayed in the evidence store, and every reload
re-materialised them. That is why the ~9k iproute2 `ss` false positives
kept "coming back" after the hash was allowlisted a day earlier -- nothing
retracted them, and the only cleanup was `doctor fix
retire-allowlisted-findings`, which operators do not know to run.
allowlist-add now retires the now-allowlisted findings at the moment of
the action (alongside the existing containment release) and reports the
count as `retired_findings` in its reply. retire_allowlisted_findings is
hash-based and self-filtering, so a path/rule entry that matches no
fully-allowlisted finding retires nothing.
Regression test (test_daemon_runtime): a detected binary's finding is
materialised (forced correlation), allowlist-add reports
retired_findings >= 1, and after correlation re-runs the finding does not
come back (A already makes correlation skip allowlisted entities).
Also refresh the doctor findings.allowlisted message, which claimed
allowlisting does not delete stored findings -- it now does.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
On a host that sits at critical/emergency for a long time (its own apps holding
most of RAM, or the daemon slightly over its soft budget), the starvation guard
was disabled (`!hard`), so correlation, retro-hunt and inventory deferred
forever — observed deferred 153 min with forced=0, which also starves the
retro-hunt delayed labels that Stage 4 training needs.
- Tiered starvation guard: it now fires under critical/emergency too, but after a
level-scaled window (1x warning, 4x critical, 12x emergency) so background work
is rare under load, not zero. A true host-memory crunch (host_severe_,
MemAvailable below the severe floor) still holds everything, since forcing work
then could tip the host toward OOM. exec guard/scan paths unchanged.
- Raise the balanced own-memory soft budget 512 -> 768 MiB: the daemon's steady
footprint grew (telemetry buffers, the ingest server, the shadow model), so 512
left it permanently over budget -> pinned critical/emergency. 768 fits the real
baseline, still well under the 1.5 GiB cgroup cap.
Tests updated for the 768 budget and the tiered guard (critical forces after the
4x window; severe holds it). Suite 51/51; analyzer + clang-tidy clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The teacher-label join made a window a positive if it was detected
(replay.cpp), but it never applied the allowlist, so a window running a cleared
false positive (e.g. /usr/bin/ss) was labelled a teacher positive — poisoning
Stage 4 training. Negative labels were written to the stream but replay dropped
them.
- replay_store now collects negative labels into ReplayOptions.negative_labels
(it used to ignore any non-positive label).
- A window whose exec hash is negatively labelled is marked `suppressed`; the
example/policy-eval positive is gated on `!suppressed`, so such a window is
kept as a *negative* training example (useful signal) and never a positive,
even if a cleared rule "detected" it.
- DetectionPipeline seeds a negative label for every allowlisted sha256 at init,
so evidence stored before the allowlist (the historical ss storm) is covered
without waiting for a re-exec.
Test: a detected window with a stored negative label for its hash is a positive
before the label and a kept-negative after. Suite 51/51; analyzer + clang-tidy
clean on replay.cpp and detection_pipeline.cpp.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Correlation runs over the whole store and never consulted the allowlist, so an
allowlisted hash (e.g. the /usr/bin/ss feed false positive the operator cleared)
regenerated the same findings on every pass from already-stored indicators —
`doctor fix retire-allowlisted-findings` cleared them but they came straight back
(~9,400 findings observed again). This also meant the findings were a poisoned
teacher-label source for training.
- classify() now skips an indicator whose (hash, path, rule) is allowlisted, the
single chokepoint shared by CorrelationEngine::build_findings (the oracle) and
IncrementalCorrelator::update, so both agree. New AllowlistFn; build_findings /
correlate take an optional predicate (default = no filtering, existing tests and
the incremental==full oracle unchanged); IncrementalCorrelator::set_allowlist.
- DetectionPipeline wires the predicate to PolicyStore::match, so the live
incremental path skips allowlisted evidence. Retiring the findings now sticks.
Test: build_findings with an allowlist predicate emits nothing for the matched
hash and is unaffected for others. Suite 51/51; analyzer + clang-tidy clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Completes the tracker-attribution transport on macOS. There is no network
sandbox to work around there, so the daemon (root) reads Little Snitch's traffic
stream directly and feeds the same on_opensnitch_connection consumer — no relay
or ingest socket (those are the Linux sandbox answer).
- LittleSnitchParser (cross-platform, unit-tested): parses `littlesnitch
log-traffic` CSV via a header-name->index map (tolerates column reordering),
yielding the shared redacted OpenSnitchConnection for outbound rows only
(program, uid, host/ip/port, protocol 6->tcp/17->udp). Little Snitch's CSV has
no argv/env, so redaction is structural. split_csv_line handles quoted fields.
- LittleSnitchSource (macOS only): posix_spawns `littlesnitch log-traffic
--stream`, reads the pipe on a guarded thread, parses and calls back; stop()
SIGTERMs the CLI (EOF -> loop ends) then reaps it.
- Daemon: start_tracker_attribution gains an __APPLE__ branch that starts the
source; shutdown/status handle it. New config littlesnitch_cli (default CLI
path). status privacy.listener source "little-snitch".
- docs: macOS section rewritten (needs Little Snitch "Allow access via Terminal").
Test E10 covers the parser (header, outbound/inbound, deny, quoted exe, no-dest).
Linux suite 51/51; analyzer + clang-tidy clean (Linux view). The __APPLE__ code
compiles/validates on the Mac next (ssh); live spawn needs root + LS Terminal access.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
First half of the sandbox-safe tracker transport (architecture deep dive). The
daemon's old loopback TCP listener could not bind under its own sandbox
(PrivateNetwork=true + RestrictAddressFamilies=AF_UNIX). Replace it with an
AF_UNIX ingest socket the daemon *can* bind under the sandbox; a host-namespace
relay (next change) will bind the OpenSnitch port, strip argv/env, and forward
redacted tuples here.
- opensnitch_events.hpp/cpp: encode_connection / decode_connection serialize only
the redacted fields (never argv/env, which the parser already drops) as one
JSON object — the relay->daemon wire form. New TrackerIngestServer: an AF_UNIX
stream listener (Linux-only, like the TCP listener) that accepts only a root
peer (SO_PEERCRED), frames with take_json_objects, decodes and calls back;
bounded clients/object/buffer as before.
- daemon: start_tracker_attribution now opens the ingest socket
(opensnitch_ingest_socket, default /run/strawberry/trackers.sock) instead of
the in-sandbox TCP listener, so it works with the sandbox unchanged and is
simply idle until a relay connects. status privacy.listener reports the ingest
socket + relay_events; shutdown stops it. New config key opensnitch_ingest_socket.
- Tests: E8 encode/decode round-trip + redaction (no args/env on the wire);
E9 TrackerIngestServer end to end (connect, two tuples, callback, stats);
daemon-runtime configs set privacy_trackers=false (not under test).
Suite 51/51; clang static analyzer and clang-tidy clean on the changed TUs. The
relay run-mode + systemd unit + install wiring + docs follow in WS2b/WS2c.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The scan path (handle_process) and the exec-guard recording path
(record_exec_decision) had grown two near-identical repeat-aggregation maps
(repeats_ / exec_repeats_) plus three copies of the IOC-feed severity idiom.
Refactor to one implementation; no behavior change (the existing aggregation and
IOC-severity tests are the oracle and pass unchanged).
- New include/strawberry/repeat_aggregator.hpp: RepeatAggregator owns the
per-key window map, the 20-full-then-summary cadence and the aggregated count;
observe(key, now, seen_ns, pid) returns {full, at_threshold, aggregated,
emit_summary, notify, state}. Each caller keeps its own key, summary-id prefix,
summary shape and notify behaviour. Both pipeline sinks now hold one instance
(scan notifies at most hourly; exec passes notify_ns=0). status
repeats_aggregated sums both totals (atomic, lock-free read).
- ioc_hit_severity(feed): one source for the three IOC match sites (scan_image,
assess_exec, handle_process) that was the policy_->ioc_feed_severity ternary.
- Leaves the latency-critical assess_exec/decide path untouched; both sinks still
run under mutex_, off the FAN response.
Unit test tests/test_repeat_aggregator.cpp (window reset, threshold, summary and
notify cadence, per-key independence). Suite 51/51; clang static analyzer and
clang-tidy clean on detection_pipeline.cpp; fold's 7 int-overflow are the accepted
monotonic-timestamp FP class (unchanged by the move); perf-kpi PASS, no regression
(exec_decision p50 unchanged, exec_assess/db_commit within noise or faster).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Allowlisting a hash suppresses new detections but does not delete findings
already stored for it (e.g. the /usr/bin/ss false-positive storm). That cleanup
now lives in doctor instead of a one-off compaction.
- DetectionPipeline::retire_allowlisted_findings(dry_run): a finding is stale
when it references at least one hash and every hash it references (from
metadata "hashes" and file_<sha256> nodes) is allowlisted, so a finding that
also implicates a non-allowlisted binary is never dropped. Mirrors
retire_superseded_findings (retire_finding + transaction).
- node_ids.hpp gains content_hash_of_file_node_id() (the inverse of
file_node_id for the content case) so the pipeline parses file ids through the
canonical helper, not a "file_" literal (source invariant R8).
- Doctor check findings.allowlisted (warn) with maintenance action
retire-allowlisted-findings; DoctorInputs.allowlisted_findings gathered by the
daemon; run_maintenance handles the action; CLI `doctor fix` help + completion
and API_CONTRACT updated.
Tests: test_retire_allowlisted_findings (two execs of a flagged binary -> a
finding; allowlist the hash -> dry-run lists it, apply retires it, idempotent);
doctor check ok/warn cases; CLI completion string. Suite 50/50; clang static
analyzer, clang-tidy and source invariants clean on the changed TUs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
The exec-guard would-block recording path stored one indicator (and, via
correlation's per-process grouping, one finding) per exec, with no aggregation,
unlike the scan path. A critical feed or rule matching a frequently-run binary
would flood evidence the way /usr/bin/ss did.
- record_exec_decision now collapses repeated identical decisions (keyed by
verdict|hash|rule) with the same 10-min window / 20-full-then-summary scheme
as handle_process (new exec_repeats_ map; counts into repeats_aggregated_).
Beyond the window it refreshes one "exec-repeat" summary indicator and returns
before the per-exec graph write.
- The actor's /proc identity reads are now done before taking mutex_, not under
it. The clang static analyzer (new strict gate) flagged ::read in a critical
section (unix.BlockInCriticalSection) here; this also honors the AGENTS rule
that scans/reads stay off DetectionPipeline::mutex_. Analyzer now clean.
Test: test_repeated_would_block_is_summarized drives 30 identical would-block
decisions and asserts exactly 20 full exec-audit indicators + 1 exec-repeat
summary and repeats_aggregated == 10. Suite 50/50; analyzer and clang-tidy clean.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
A benign system binary (/usr/bin/ss) whose SHA-256 was on a noisy community hash
feed produced thousands of critical "would have been blocked" findings: IOC hash
matches were hardcoded severity "critical" at every site, and in exec_guard an
ioc_match forces a prevention hit regardless of the severity threshold.
- FeedSpec gains `max_severity` (append-only; serialize/parse/validate, control
socket `feed-add` accepts it). Empty = default for the kind: hash feeds are
advisory "high", not "critical". PolicyStore::ioc_feed_severity() resolves it.
- All three IOC severity sites now use the feed's severity
(detection_pipeline.cpp scan_image, assess_exec, handle_process), replacing the
hardcoded "critical".
- Trust gating: ExecAssessment::ioc_match is set only when the matched feed
resolves to "critical" (a curated feed that should force prevention). A
community feed stays advisory — its "high" is subject to prevention_min_severity
(default "critical"), so it no longer would-blocks system tools per exec.
Tests: test_ioc_hash_detection uses an explicit critical feed (trusted path) and
asserts assess_exec forces the hit; new test_ioc_community_feed_is_advisory
asserts a default feed is "high", ioc_match is not forced, and decide_exec allows
it in audit and enforce; test_policy covers max_severity validation, the advisory
default, and a critical round-trip. Suite 50/50; clang static analyzer and
clang-tidy clean on the changed TUs.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Second piece of tracker attribution (docs/tracker-attribution.md): which
programs reach domains on a trackers list. Advisory only: never a block,
never a finding (source invariant R11).
- OpenSnitch event listener: OpenSnitch's JSON "remote" logger is the
hostname source (Strawberry's eBPF connects have none). It streams over
TCP and cannot leave fields out, so events carry command lines and
environments. The listener binds 127.0.0.1 at a port below 1024 (only
root can), accepts only root-owned peers (/proc/net/tcp), bounds
clients/object size/buffering, and reads only pid, uid, path,
destination and action; arguments and environment are never copied.
- TrackerContacts: per program x domain, bounded at 20000 keys. A root-only
tracker_contact telemetry record (type 25) on the first contact, then at
most hourly with the count since.
- Status `privacy` block; control verb privacy-trackers; CLI
`strawberry privacy trackers [--json]`.
- `strawberry opensnitch enable-events` adds the logger to OpenSnitch's
config (with a .bak); the installer runs it when OpenSnitch is present,
unless privacy_trackers=false. Config: privacy_trackers (default true),
opensnitch_events (default 127.0.0.1:979).
- EasyPrivacy is a default trackers feed (GPL-3.0+ or CC-BY-SA-3.0,
downloaded at run time).
Tests: E1-E7 (test_opensnitch_events: event fields, stream splitting,
peer uid, address rules, listener end to end, contacts and record, config
editor), the control verb, CLI translation and help, daemon status. New
fuzz_opensnitch_events (90 s clean).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
First piece of tracker attribution (docs/tracker-attribution.md): which
programs reach advertising and tracking domains, built with the
measurement rules of the CarolinaCon 2026 privacy talk (bridgescope):
a listed domain is a circumstantial label, never proof.
- Feed kind "trackers": hosts lines, plain domains and adblock
whole-domain rules (||host^, ||host^$third-party). Cosmetic rules,
exceptions, paths and other options are skipped; the comment cut used
to turn "site.example##.ad" into a domain.
- TrackerSet: suffix matching (a.b.t.example matches t.example, never
nott.example, an IP literal or a bare TLD); first feed reported.
- List parsing moves from rule_updater.cpp to ioc_list.cpp (no network
code) so the query-side core can use it.
- Source invariant R11: tracker lists never meet exec-guard or OpenSnitch
code. They are not in the OpenSnitch blocklist sync.
Tests: T1-T3 (test_tracker_set), trackers in the list-parsing property
and the list fuzzer (60 s clean), R11 self-test fixture.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
`model train` on a host without detections stops with "0 positive
windows" and pointed at --inject-train FILE, but only the evaluation set
(standard.trace) existed. Add config/traces/training.trace: 7 attack
shapes (document macro dropper, web reverse shell, cron miner, /dev/shm
perl stager, ssh key backdoor, npm postinstall, ld.so.preload) that differ
from the evaluation set, so the gate never judges a model on what it
trained on. P4 in test_trace enforces that (no shared trace name or exec
sequence); it caught an evince -> python3 overlap in the first draft.
Both installers put the traces in /usr/local/share/strawberry/traces/
(install_traces in install_common.sh); the hint and README name it.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The shadow scorer's rarity model started empty at every daemon start, so
each restart cost an hour of warm-up with no decisions. The daemon now
saves it to model_dir/rarity.state (checksummed, atomic write, hourly and
at stop) and restores it at start. A restored state younger than the
7-day half-life skips the warm-up; decay catches up on the first event.
A file made with other rarity settings, a bad checksum, or a negative or
non-finite count is refused with a reason (status model.rarity_state,
doctor model.rarity_state) and the model starts empty. The file must be
owned by the daemon user and not group/world writable, like model files.
Also fixes status model.warmup_windows, which was never filled in.
Tests: R1/R2 in test_learn (round trip, decay clock, refusals), S8 in
test_shadow (warm-up skipped only when recent), daemon restart restores
the saved state, doctor warning; new fuzz_rarity_state target.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
A promoted model (promoted.sbm, written only by `model promote` after the
gate) can now request real inspections. Only two, both additive:
- memory_scan: through the daemon's memory-scan queue;
- deep_rescan: the root's image re-scanned on demand; new matches are
recorded as indicators of kind model_inspection (findings titled "YARA
match found by a deeper rescan the inspection model requested") and
positive labels.
watch_subtree has no mechanism and rescan_on_next_rules is the retro-hunt's
job, so both are masked out of acting decisions (kExecutableActions), and
replay's policy evaluation uses the same action set.
Guardrails:
- InspectionExecutor (bounded queue, one low-priority worker, submit()
never blocks); WorkClass::model_inspect lets the governor defer it, and
a deferred decision is dropped.
- Each action rechecks that the promoted model is still active at the
decision's generation (demoting or replacing it revokes queued work),
that the root's node id and raw image path still match (decisions carry
root_image, not the feature template), and skips protected processes.
A deep rescan discards its result if the process, image or rules changed
during the scan.
- An acting decision is durably written before it is dispatched; every
attempt becomes an inspection_outcome record (queued, scanned, skipped,
deferred, revoked, queue_full, failed).
- Source invariant R10: no file may couple learning types to the
exec-guard, OpenSnitch or policy APIs (daemon_app.cpp and main.cpp, the
wiring points, are reviewed by hand). Models never block, contain, kill
or create findings directly.
- Config model_act and model_lambda; status model block reports acting
stats; timings.model_inspection.
docs/evaluation/stage4-execution-gate.md amends the gate's method (not its
thresholds): executable action set, greedy promoted decisions, budgets as
estimated inspection seconds by scan profile.
Also: TerminalSafeBuf's destructor no longer lets an exception escape;
the health message clamps huge durations; replay timestamp overflow; the
Docker test build runs with -j2.
Tests: test_inspection_executor (every skip and execute path, revocation,
queue limits); shadow and replay acting cases; R10 fixture; fault injection
and invariant updates. Validation reported by the session that wrote it:
48/48 local, 46/46 Docker plus privileged live checks, 20 fuzz targets
(9.5M runs), three paired KPI runs without regressions (no acting model
was loaded, so active-inspection cost is not measured yet).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- replay --policy-eval: over ranked windows, simulates the model and
rarity-greedy (the same chooser, rarity score / 48 as the value), each
with its own hourly budget, and tallies the shadow decisions the daemon
logged. Hits are inspected "extra" positives: late labels or injected
evaluation traces on windows rules did not flag at exec time. Labels are
known after the fact, so every action's payoff is computable without
propensity weighting. Reports CPU-seconds, inspections, hits, recall,
hits per CPU-second and budget refusals.
- docs/evaluation/stage4-promotion-gate.md (pre-registered): at least 10
extra positives, model efficiency >= 1.5x rarity-greedy, model recall
not lower. promotion_gate() checks it.
- strawberry model promote [--inject FILE] [--warmup H] [--yes]: replays
with the shadow model, applies the gate and only then writes
promoted.sbm; asks for confirmation.
- learn/inspection.hpp: options, the greedy chooser and the
inspection_action record, shared by shadow mode and the evaluation.
- Fix: the hourly budget reset whenever a window closed with an earlier
timestamp than the hour's start (windows close roughly, not strictly,
in time order), letting spend exceed the budget; in both shadow mode
and the evaluation.
Tests: test_replay R11 (gate criteria; extra positives exclude rule-flagged
windows; spend within the budget; logged shadow decisions joined to their
windows); CLI translations.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The daemon now scores live windows with the shadow model and logs the
inspection it would request. Nothing is executed.
- learn/shadow: a live replay engine (same windows and rarity as
`strawberry replay`) fed every telemetry record on the writer thread.
When a window closes, the model's value v gives each action a benefit
v x reach - lambda x CPU cost; the best positive one is chosen if the
hour's budget (10/30/120 CPU-s by scan profile) allows, else none.
With probability 0.02 a uniformly random other action is taken instead,
and every decision records its propensity for off-policy evaluation.
Decisions are inspection_action telemetry records.
- Replay gains a live mode (no ranking bookkeeping, so memory stays
bounded) and a window-closed callback.
- Daemon: model_shadow (default on) loads <model_dir>/shadow.sbm when it
appears or changes, checked every minute; an invalid file is reported
and not used. Status gains a `model` block (kernel, generation,
decisions by action, exploration, would-be spend, budget refusals).
- learn_sources in src/meson.build: tests that compile the daemon from
sources take the learning library whole.
Limits, documented: decisions come when a window closes (a few minutes
after its first exec), and the live rarity model starts empty at each
daemon start (the first hour is learned, not decided).
Tests: test_shadow (no model, greedy choices, budget never exceeded,
determinism, exploration rate and propensities, record codec); daemon
status; fuzz_telemetry_records covers inspection_action.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Replay builds each window's WindowSummary as it goes (exec depth and
locations, connections, writes, ptrace, module loads, scan outcomes,
surprisal) and can collect training examples: features, a label (a
detection or late label touched the window), time, root program and the
rarity score. Sampled to at most 50k: positives up to a fifth,
negatives by reservoir with at most 1% per root program.
--inject-train traces become positive examples and are not evaluated.
- learn/train: the value head (and hidden layers) learn P(teacher
positive) with a class-weighted log loss and Adam; positive-unlabeled
data, so unlabeled is "not known bad", documented. Split by time
(newest 20% tested); reports test loss, top-k hits and ROC AUC for the
model and for the rarity baseline. Refuses with fewer than 5 positives.
Deterministic for a seed. The policy head waits for shadow-mode
inspection outcomes.
- CLI: `strawberry model train [store_dir] [--inject-train FILE]
[--epochs N] [--seed N] [--out FILE] [--json]` writes the shadow model
(generation after the newest existing one, a note saying what it was
trained on); `strawberry model show [file]`; `strawberry replay --model
FILE` ranks windows (and evaluates injections) by the model's value.
Config model_dir (default /var/lib/strawberry/model).
- doctor: the model check loads promoted.sbm and shadow.sbm; an invalid
file is a problem (it is never used).
Tests: test_train (gradients against finite differences, skipping steps
that cross a ReLU kink; a planted pattern learned to high AUC; determinism;
refusal without labels; roc_auc); test_replay R10 (examples, labels,
per-program cap, training traces, model ranking); CLI and doctor cases.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Clang rejects "\x80evil" (the escape takes the 'e'); GCC accepted it and
the test used a different string than written.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
New meson option enable_endpoint_security (auto, macOS only) links
EndpointSecurity and bsm and builds three ES backends behind the
existing interfaces. Helpers are in src/platform/es_support.hpp.
- ProcessMonitor: NOTIFY_EXEC events (pid, parent, path, dev:ino,
argv, start time) plus NOTIFY_TRACE and NOTIFY_KEXTLOAD for the
behavior rules. Events go to a bounded queue (4096, overflow counted
in events_dropped) drained on the monitor thread, so callbacks stay
on one thread. Without an ES client the monitor polls libproc
(backend "libproc"). backend_name() reports "endpoint_security".
- ExecGuard: AUTH_EXEC with the fanotify guard's rules. It fails open
at min(timeout, ES deadline - 500 ms) and on a full queue, and on
macOS 27+ ES_DEADLINE_MISS_MODE_FAIL_OPEN makes the kernel allow a
missed deadline too. Platform binaries and SIP paths are never
denied, and our own execs are muted. ES does not cache verdicts, so
rule changes apply on the next exec. Our own cache keys on
dev:ino:size:mtime:generation. The image is opened by path and its
dev:ino checked against the event; a swapped file counts as
unscanned.
- FileWatcher: NOTIFY_CREATE/RENAME/CLOSE(modified) with inverted
target-path muting: named entries selected literally, whole
directories by prefix. macos_watch_specs() covers launchd
agents/daemons, periodic jobs, cron tabs, sudo drop-ins, shell rc
files, authorized_keys, Downloads and the temp directories.
match_watch_spec() (direct children, as with fanotify directory
marks) is portable and tested. Home dirs come from the directory
service (local_home_dirs).
Every ES handler is noexcept and records swallowed exceptions. Client
errors name the missing requirement (root, entitlement, Full Disk
Access), and the daemon keeps running detection-only.
macos_es_live (root, ES entitlement) checks deny/allow/cache/
platform-binary/fail-open/audit for the guard, exec telemetry and
the file watcher, using small signed payload binaries. It has not run
yet: the development Mac has SIP enabled.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
include/strawberry/platform_process.hpp answers what /proc answers on
Linux: running PIDs (proc_listallpids), executable path
(proc_pidpath), parent and start time (sysctl KERN_PROC_PID, readable
for every process), argv (KERN_PROCARGS2) and a boot id
(kern.bootsessionuuid). A process is identified by its start time in
microseconds since the epoch plus the boot id, so node ids survive PID
reuse as on Linux. It is header-only, so the per-test source lists stay
as they are.
Users on macOS: resolve_process_ref (default root), the pipeline's
image hashing and exec-decision actor, and YaraScanner::scan_process
(scans the image by path). A synthetic procfs root still takes the
procfs code, so the existing procfs tests run on macOS too.
The resource governor samples Mach/sysctl on macOS: own CPU from
getrusage, RSS from task_info, load from getloadavg, available memory
from host_statistics64 (free + inactive + purgeable + speculative).
The behavior engine knows macOS debuggers (debugserver, dtrace) and
kext loaders (kernelmanagerd, kmutil), and macOS temp and Downloads
paths count as drop locations. The Linux lists are unchanged.
The KERN_PROCARGS2 parser is portable: a property test (exact argv
recovery, truncation never invents data, total on junk) and the
fuzz_procargs target run on every platform. test_platform_process
checks the live helpers against this process; the daemon test reads
the stopped state through sysctl on macOS.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The tree now compiles on macOS. Each Linux-only call gets a Darwin
branch, and the Linux branch stays as it was:
- io::fdatasync uses fcntl(F_FULLFSYNC): fsync alone only reaches the
drive cache on APFS, and the store's crash atomicity relies on it.
- Control socket: FD_CLOEXEC set after socket/accept/recvmsg (no
SOCK_CLOEXEC, accept4 or MSG_CMSG_CLOEXEC), SO_NOSIGPIPE, peer
credentials from LOCAL_PEERCRED + LOCAL_PEERPID, getgrouplist with
int groups and a growing buffer. launchd has no RuntimeDirectory, so
the socket directory is created (0750, access group) when missing.
Default socket path is /var/run/strawberry/control.sock.
- CRC32C detects ARMv8 CRC with hw.optional.armv8_crc32 (no auxv).
- stat mtime via st_mtimespec; the BPF shared header uses stdint
typedefs off Linux.
- Background threads join the PRIO_DARWIN_BG band (lower CPU and I/O);
crash safety zeroes RLIMIT_CORE in place of PR_SET_DUMPABLE.
- The desktop app drops Qt DBus off Linux and shows notifications as
tray messages (app/notifier_tray.cpp).
- make setup passes Homebrew's keg-only libarchive to meson.
Tests: behavior_live (Linux ptrace API) builds on Linux only, the
control-server fd count reads /dev/fd, the crash test checks the core
limit on macOS, and the runtime pipeline test no longer references a
YARA-only helper when YARA is absent (that also broke Linux builds
without libyara).
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Nothing runs in the daemon yet; these are the pieces the next increments
train and run in shadow.
- learn/features: a WindowSummary accumulated from a window's events
(exec locations, memfd/deleted/uid0 execs, depth, connection classes
and ports, persistence and drop writes, ptrace, module loads, scan
outcomes, capped surprisal per tuple kind, bounded tokens) and
featurize(): 48 explicit features scaled to about [0, 1] plus signed
hashing of tokens into 208 dims, with the governor level and budget
left as context. A window's own detections are never features: they
are the teacher labels.
- learn/mlp: 256 -> 128 -> 64 ReLU with a policy head over the
inspection actions (none, memory_scan, deep_rescan, watch_subtree,
rescan_on_next_rules) and a value head. Scalar reference, AVX2+FMA
chosen at runtime, NEON on aarch64; safe default (no inspection) on a
wrong input size or an inconsistent model. Deterministic init per seed.
- learn/model_file: "SBMD" format with bounded dimensions, exact length,
a sha256 trailer and finite weights checked on load, each failure
named; atomic save into a 0700 directory.
Tests: test_model (kernels against scalar for arbitrary sizes, forward
pass invariants, deterministic init, round-trip, every flipped byte and
truncation rejected, non-finite weights rejected, features bounded and
deterministic). fuzz_model_file. NEON is built and tested only on aarch64.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Decided 2026-10-27 (docs/cli-ux.md):
- Exit codes: 0 done/healthy, 1 negative result (unhealthy, not found,
daemon refused), 2 usage, 3 could not run here (daemon unreachable,
permission denied, store not readable, feature not built). Daemon
unreachable used to exit 2, like a usage error.
- Process kill/suspend, quarantine, allowlist-add, protection-set to
enforce, store compact and doctor fix (without --dry-run) say what they
will do and ask "Proceed? [y/N]" on a terminal; Enter or EOF is no.
Without a terminal they need --yes, else exit 2 with nothing done.
This applies to the flag form too (--policy-set process-kill ...).
The desktop app talks to the control socket directly and is unaffected.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- Evidence strings are attacker-controlled: every CLI command's stdout
and stderr now pass through a filter that shows control characters
(C0, DEL, C1) and invalid UTF-8 as \xNN (src/term_safe.hpp), so a
crafted process name or path cannot inject terminal escape sequences.
The daemon's journal output is unchanged.
- `strawberry` alone prints help; an unknown word (`strawberry fix`) is an
error with a "did you mean" hint. Both used to start a daemon as the
calling user with the word as its config path. Config paths (contain
'/' or end in .conf) still run the daemon, for older units.
- Shell completions generated from the command table: `strawberry
completion bash|zsh|fish`; install.sh installs them under
/usr/local/share.
- One correlation pass at daemon start (stored findings catch up with the
current rules, e.g. stable ids), and retire-superseded-findings
correlates first, so replacements exist before old copies are retired.
The user's first run reported "no duplicates" for that reason.
- docs/cli-ux.md: CLI UX rules adapted from yams (grammar, flags, human
and JSON output, status words, security rules for output, help and
completions, adding a command, known gaps). Exit code 3 and
confirmation prompts are proposed there, pending a decision. AGENTS.md
and CLAUDE.md point to it.
tests/test_cli.cpp: bare and unknown words, terminal_safe and the stream
filter, completion scripts.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
strawberry [--config PATH] [--json] <command> [<subcommand>] [args] [--flags]
status, doctor, doctor fix <action>, daemon run, findings list/show/
correlate, process search/trace/trace-node/suspend/resume/kill, scan
file/processes/memory, rules reload/update/check/fetch, policy show/set,
opensnitch enable-checksums, store verify/compact, replay, demo seed,
help [command], version.
One table in src/cli_commands.cpp defines each command (words,
positionals, flags, help) and how it maps onto the existing flag form,
which main() still parses; help is generated from the same table.
Unknown options and wrong arity print the command's usage. Process
actions validate node ids before building their JSON payload.
The older flag form (strawberry --status, strawberry <config>) still
works. The systemd units now run `strawberry daemon run --config ...` and
`strawberry rules fetch --config ...`. README, AGENTS.md and the
false-positive guide use the command form.
doctor renders the daemon's checks with their fixes; doctor fix runs a
maintenance action (with --dry-run). Against an older daemon the CLI says
to reinstall rather than "unknown command". replay explains that
telemetry is root-only instead of reporting zeros, and keeps config
output out of --json.
tests/test_cli.cpp: translations, flag-form passthrough, errors, help.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- doctor (control verb, read-only): store health and disk floor,
telemetry growth per day and days until the floor when retention is
unlimited, leftover store files, superseded findings, retro-hunt state,
thread/write errors, logs (journald rotates them), model (none yet).
Each check carries a maintenance action or advice. Pure checks over
gathered inputs (src/doctor.cpp), so they test without a daemon.
- maintain (control payload verb, privileged, rate limited, audited):
retire-superseded-findings and clean-store, each with dry_run.
- finding_retired evidence record: removes a finding from view; it stays
in the stream's history. CorrelationEngine::superseded_findings finds
old hash/rule findings by recomputing the stable id, and old lineage
findings only when the current one's parent process is among their
nodes (a reused pid on another boot is not a match). Lineage findings
now record parent_process_node_id.
- store::inspect_store_files / clean_store_leftovers: orphan/ files older
than N days and *.compact / MANIFEST.tmp temp files; never a listed
segment.
Tests: test_doctor (check statuses, superseded detection including pid
reuse and no-replacement cases, retirement surviving a reopen, cleanup
keeping recent orphans and listed segments); daemon test exercises both
verbs over the socket. fuzz_evidence_records covers finding_retired.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The main source of teacher labels over time: rules improve daily, while
most malware runs once.
- The pipeline remembers every binary that ran (image_seen evidence
record: sha256, path, first/last seen, size; at most one write per hash
a day). Deleted and memfd images are skipped: nothing to find again.
- Each rules load computes a digest of the rule files and IOC hash lists
(paths after resolving symlinks, sizes, mtimes).
- RetroHunter (src/retro_hunt.cpp, own thread, idle priority, paced at
retro_hunt_pace_ms, deferred by the governor as WorkClass::retro_hunt)
re-scans when the digest differs from the last completed pass, most
recently seen first. Only paths that still hold the same content are
scanned (O_NOFOLLOW, regular files, hash compared); others count as
changed or missing. A completed pass is recorded (retro_hunt record) so
a restart does not repeat it; an interrupted one starts over.
- New matches (rules not recorded before for that hash, not allowlisted)
become indicators with scan_kind retro_hunt, a file/indicator graph
link, a positive label and a notification. Correlation turns them into
findings titled "YARA match on a binary that ran here (found after a
rules update)". Nothing is blocked or contained from here.
- Replay reads positive labels from the evidence stream and marks windows
that ran a binary labelled bad later (LATE in the CLI).
- Status: a retro_hunt block (API_CONTRACT.md); config: retro_hunt,
retro_hunt_pace_ms.
Tests: test_retro_hunt (remembered once; passes only on a digest change
and survive a restart; a clean-at-exec binary matching a new rule becomes
an indicator, label, finding and notification; a replaced binary is
counted as changed, never scanned as the old one; a stop abandons a pass
unrecorded); test_replay R9 (late labels, from a real store).
fuzz_evidence_records covers the new records.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
- src/learn/trace.cpp: a small trace format (exec, connect, write,
ptrace, module with +offsets and process aliases) compiled to
telemetry records with pids above any pid_max and node ids that name
the trace and copy.
- --replay --inject FILE [--inject-copies N]: copies are spread evenly
over the ranked history (a first pass reads the time range), scored
but never learned from, and checked against alert budgets of 1, 3 and
10 per day: is the copy's best window above the (budget x days)-th best
real window? Ties count against. Injection leaves real windows and
their ranking unchanged.
- config/traces/standard.trace: the six plan traces (browser to curl|sh,
systemd user persistence, memfd loader, ptrace injection, download
then exec, python from a PDF viewer).
- docs/evaluation/stage3-rarity-preregistration.md: method and pass
thresholds, written before any run on host data.
Tests: test_trace (standard traces, error messages with line numbers,
records); test_replay R8 (every copy found, real ranking unchanged, a
novel chain caught at 10/day and a bare shell not). fuzz_trace_dsl.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
On the host, 586 of 303k records arrived more than 30 s late: exec records
of large images are written after their paced, deferrable scan. Replay is
offline, so holding 2 minutes of records costs nothing that matters.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
From the first replay of real host data (3.5 h, 266k records):
- A never-seen event scored higher the more the model had seen, so the
ranking favoured whatever happened last (the reinstall's systemctl
and generator runs). Each event is now capped at 16 bits (never
seen), and a window scores the sum of its three most surprising
distinct events.
- Warm-up: on a fresh history nearly everything is new. Windows in the
first 12 h are learned from but not ranked (--warmup H); detections
are always listed. The CLI says so when nothing is ranked yet.
- Windows with the same root program and top pattern are one line with
a count ("x40 similar").
- Counts: file writes split by persistence; events from processes
started before recording are reported as such, not as an error;
records more than 30 s out of order are counted as late.
tests/test_replay.cpp R7: warm-up, equal scores for a novel event early
and late, grouping, late-record counting.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Replay (src/learn/replay.cpp, strawberry --replay [dir] [--top N] [--json]):
reads the telemetry stream read-only, orders events by time (a 30 s
reorder buffer absorbs the writer's batching), joins behavior events to
the latest exec of their pid at that time (keeps PID reuse apart), builds
60 s windows (a root exec, its descendants and their behavior, at most 64
events), scores each exec, connection and persistence write with the
rarity model before observing it, and marks windows whose processes had
detections. Prints the most unusual windows with the tuples that made
them unusual. Deterministic. tests/test_replay.cpp: chain windows, PID
reuse, order insensitivity within the horizon, determinism, ranking,
detection marking, a real store round trip. fuzz_replay feeds arbitrary
record streams.
UTF-8: file names are bytes, and the fuzzer showed nlohmann's dump()
throwing on a path that is not UTF-8. The same applied to the daemon's
control replies (kill/suspend/quarantine name the exe path) and to the
OpenSnitch containment rule, so a binary with such a name could not be
contained, and json_escape wrote the raw bytes into status JSON. Now:
- every .dump() in src/ uses error_handler_t::replace (U+FFFD), enforced
by source invariant R9;
- json_escape writes invalid sequences as � (valid_utf8 and
utf8_sequence_length in json_util.hpp);
- containment refuses such a path with a clear error (a rule could never
match the real bytes); quarantine still blocks execution by hash.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
src/learn/ (pure libraries; nothing in the daemon calls them yet, Stage 3
uses them in replay):
- tokenize: path templates (user homes, per-uid dirs, pids, digit and hex
runs, UUIDs, mktemp-style names abstracted) and command-line shapes
(flags kept, values, paths, URLs, IPs, numbers and encoded blobs
classed; the first plain word kept as a subcommand; long argument lists
summarized). Bounded and noexcept.
- rarity: decayed count-min sketches over parent>child, grandparent>
parent>child, exe+command-line shape, exe+connection class and
exe+persistence location, scored as surprisal in bits; a bounded
first-seen table; deterministic hashing for replay. About 1.3 MiB.
tests/test_learn.cpp: tokenizer cases and idempotence; count-min never
underestimates (also after decay); determinism; rare scores above common;
half-life decay; bounded first-seen. fuzz_tokenizer found two
idempotence bugs (truncation before templating, and a cut that left a
" (deleted)" suffix), fixed here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Their ids included the member pids (and rules or hashes), so every new
process joining a group created a new finding and left the previous one
behind as a duplicate. The id is now the group key alone: the hash, the
rule name, or the parent process. The finding is updated in place as the
group grows; the pids stay in its metadata and summary.
Findings stored under the old ids stay as they are (one last duplicate
per group after upgrading). The incremental property test now also checks
that add-only runs leave exactly one finding per group.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Every finding is a function of one group of indicators (a process, a
parent, a hash or a rule). IncrementalCorrelator keeps group membership
(ids and keys, not indicator contents) and rebuilds only the groups that
new or changed indicators touch. Each pass used to copy every indicator
and rebuild every finding; it now costs about the size of what changed.
- correlation_engine.cpp: the four group loops become per-group
functions (re-indented; logic unchanged), shared by the full pass and
the incremental one.
- EvidenceView keeps a bounded change log of written indicator ids
(resets after clear(), at first use, or past 65536 entries);
Database::take_indicator_changes / find_indicators expose it, with
rolled-back writes reported as removed.
- DetectionPipeline holds its lock only to read changes and members and
to store findings, as before.
Fix found by the oracle test: an indicator naming a process node but no
pid formed a second group for the same process with the same finding id,
and the two findings overwrote each other. The group pid now comes from
the node id when the pid is missing (finding ids are unchanged).
tests/test_correlation_incremental.cpp: random batches of new and updated
indicators give exactly the findings the full pass after every batch
stored (I1, I2); a batch touching one process rebuilds only its groups
(I3); the store's change log, including a rolled-back write.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
File node ids: include/strawberry/node_ids.hpp is now the only place
they are built (file_<sha256> or file_path_<sha256 of path>). The three
modules disagreed: correlation used a 16-digit path hash and behavior
detections the raw path, so path-keyed evidence never met the rest. An
exec detection with a known hash also writes a same_path edge from the
path id to the content id. Source invariant R8 rejects hand-built file
ids in src/ and include/.
Processes carry the real uid and kernel comm: from the eBPF exec event
(which now includes the uid) or /proc. Stored on process nodes and in
process_exec telemetry (fields 14, 15).
Command lines are redacted before any storage (capture_redact_cmdline,
default on): sensitive flags and KEY=value pairs, Authorization/Bearer
credentials, URL passwords, tokens used as URL user names, well-known
token shapes (GitHub, GitLab, Slack, AWS key ids, sk-, JWTs) and mysql
-p<password>. The rules repeat to a fixed point, so redaction is
idempotent by construction. tests/test_redact.cpp covers the rules plus
idempotence, size and planted-secret properties; fuzz_cmdline_redactor
found two idempotence bugs and a leak (--password --token X) that are
fixed here.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Teacher signals and context for the learning stages, written to the
evidence stream (retention never removes it):
- label: operator actions from the control socket (quarantine, contain,
kill, suspend: positive; allowlist-add: negative; allowlist-remove and
quarantine-remove: retracted) and allowlist suppressions (negative,
once per entry, subject and rule). Never from the MCP.
- rules_generation: every rules load attempt, with sources, counts and
the error of a failed reload.
- policy_generation: each policy change, with a sha256 of the policy.
- lifecycle: daemon start and stop.
governor_sample (pressure level, CPU/IO/memory, per-class admissions) goes
to the telemetry stream once a minute.
Database::append_evidence writes records the evidence view ignores.
Tests: record round-trips; suppression labels and dedupe and rules
generations in the pipeline test; operator labels, policy generations and
lifecycle in the daemon test; the record fuzzer covers the new decoders.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Each rule match from a fresh scan now writes a rule_match record: the
rule, namespace, severity, rules generation and its matched strings
(identifier, offset, length, 64-bit FNV-1a digest of the captured bytes).
The matched bytes themselves are never stored. Cache hits write no
rule_match, since their scan_result already names the rules.
Hits are collected in the libyara callback (libyara 4+) and capped at
16 strings per rule, 4 hits per string and 256 per scan. The callback
also tolerates a null user_data now.
Tests: caps, offsets and digests against a real scan; record round-trip
with the encode cap; the telemetry record fuzzer covers rule_match.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Stage 2 increment 2.
- Connect, ptrace and module-load events go to telemetry raw, before the
behavior engine filters them (pid, uid, comm, address/port or target);
the join to processes happens when the data is read.
- Persistence and drop-location writes go to telemetry before any rule
decides on them.
- The exec guard reports every verdict through a new on_observation hook
(allow verdicts and cache hits included; ExecDecision still reports
only non-allow): verdict, answered, scanned, cache hit, degraded, top
rule/severity and latency, as guard_decision records.
Tests: record round-trips; the live exec-guard test checks one
observation per checked exec, including allows and cache hits;
fuzz_telemetry_records covers the new decoders.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Stage 2 increment 1. Until now only detections were stored, so the
learning stages would never see a normal example.
- TelemetrySink: bounded lock-free multi-producer queue (8192); try_emit
never blocks, allocates or throws; a full queue drops and counts.
- A guarded writer thread drains it every 2 s (or at a quarter full)
into the store's root-only telemetry stream through the pipeline,
writing a gap record whenever events were dropped.
- The pipeline emits one process_exec (pid, parent, node ids, exe,
capped cmdline, sha256, deleted/memfd, live/inventory) and one
scan_result (outcome: scanned, cache_hit, too_large, unavailable,
read_error, no_rules, yara_error, pid_scan; bytes, scan time, rules
generation, matched rules) for every handled exec, detection or not.
- Config telemetry=true (default); status gains a telemetry block
(emitted, dropped, queued, written, gaps).
Tests: C1 (never blocks; exactly once and in order under concurrency,
clean under ThreadSanitizer), record round-trips, C2 (one process_exec
and one scan_result per handled exec, stored in the telemetry stream);
fuzz_telemetry_records.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The KPI detect phase still dropped events after the correlation fix.
Cause: every detection kept ~12 KB of evidence in memory, 20k detections
pushed RSS past the 512 MB budget, the governor went to emergency, and in
emergency every live exec scan waited 5 s, so the queue overflowed and
stayed full (memory that is evidence never comes back by waiting).
- The governor holds live exec scans for at most 5 s per emergency
episode, then lets them run (counted as forced); background work still
waits.
- Repeated detections of the same image and rules: the first 20 per
10-minute window are stored in full, then one summary indicator per
key keeps the count and first/last pid and time (rewritten at most
every 30 s). All detections are counted; alerts repeat at most once a
minute per key. Status reports pipeline.repeats_aggregated.
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Found by the KPI detect phase (2-3 persist failures per run):
- a finding for one rule across thousands of processes listed every
related id and outgrew the 256 KiB block limit, so the correlation
commit was refused; related lists are now capped at 200 (the most
recent indicators) with the totals in metadata
- a single record up to the 1 MiB record limit may now form a block on
its own (the decoder bound only loosens)
- a refused append records its reason ("rejected: ...") instead of
logging an empty error
- the Database no longer logs every indicator, node, edge and finding it
stores (175k lines in a 4-minute run); failures are still logged
Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>