feat: implement Pillars 2-4 and fix a production false-positive in Pillar 1

Pillar 1 v1.1 (shipped earlier) flagged benign browsing bursts as
beaconing: on short captures, grid periods longer than the observed
span never complete a cycle, so significance saturates to 1 and the
score sticks at exactly 0.5. v1.2 caps the grid at a third of the
observed span, requires a minimum number of completed cycles, and
excludes protocols that are periodic by nature (NTP, DNS, mDNS,
ICMP). Validated against 300 seeds per scenario: false positives on
bursty traffic drop from the 290-300/300 that v1.1 produced to 0/300,
with every beacon-detection scenario scoring unchanged.

Add Pillar 2 (persistent/long-lived sessions, with Zeek's fragmented
TCP sessions rejoined after a 5-minute gap), Pillar 3 (DNS tunneling
via Shannon-entropy information volume per subdomain per hour; the
Jensen-Shannon divergence was measured and does not separate tunnels
from benign CDN subdomains, so it stays in the evidence at zero
weight instead of in the score), and Pillar 4 (deterministic
noisy-OR fusion across the three pillars).

The fusion formula was validated against 8 labeled real-traffic
vectors (IoT-23, CTU-Normal) with tolerance <= 1e-4. A known
limitation is kept, not hidden: a benign Philips HUE bulb's strict
periodic polling and a 21.9h session score it above three real C2
hosts. No weight or threshold was tuned to mask this; it is
documented in the architecture docs instead.

Also fixes real bugs found along the way: ingest_dir's partition
count summed file names instead of distinct dates, DuckDB errors
were counted as bad rows, /api/alerts always returned an empty list
regardless of the last analysis, and the dashboard still claimed
every pillar was unimplemented.

Restores four design corrections that were missing from
docs/architecture.md (KL to Jensen-Shannon, the persistent-session
detector, the JA4 licensing boundary, and deterministic reduction
order), and catalogs the public datasets used for validation
(IoT-23, CTU-Normal, Mendeley DTQBC) with their licenses and SHA-256
checksums under research/validation/.

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01VAyr8d7HuAdUGB6aA5xmDD

CHANGELOG.md

@@ -6,13 +6,27 @@ Format based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/); versio
 
 ## [Unreleased]
 
+### Added
+- **Pillar 1 (v1.2):** Rayleigh/Schuster periodogram with grid bounded by the observed span ($T \le \text{span}/3$), regularity gating, and infrastructure exclusions (NTP, mDNS/LLMNR/NBNS, SSDP, non-TCP/UDP).
+- **Pillar 2 (v1.0):** Persistent TCP session detector with inactivity chaining (gaps $\le 3600$s without SYN), logarithmic duration scoring, and PCR / failed attempt metrics in evidence.
+- **Pillar 3 (v1.0):** DNS subdomain information capacity detector ($I = H \times C$ bits/h) to identify high-entropy tunnels, with Jensen-Shannon Divergence (JSD) recorded in evidence.
+- **Pillar 4 (v1.0):** Per-host weighted noisy-OR fusion ($w_1=w_2=w_3=0.75$) with deterministic evidence linkage, ranking hosts in `acaua analyze` human output via the `hosts:` section.
+- **Zeek `dns.log` ingestion:** Daily Hive Parquet partitioning for DNS queries with deterministic column mapping by TSV header.
+- **Parquet storage schema (v1.2):** 15-column schema for `conn.log` supporting `conn_state`, `history`, `orig_pkts`, `resp_pkts`, `orig_ip_bytes`, `resp_ip_bytes`.
+- **In-memory alert reporting:** Web daemon `/api/alerts` route now serves alerts from the most recent analysis run.
+
 ### Changed
+- **Breaking (0.x):** Pillar 1 formula v1.2 modifies scores and changes `Alert.id` generation to reflect the corrected bounded grid and session-based aggregation.
 - Distribution pivoted to 100% Docker and docker-compose (server daemon and web dashboard); native binary tarballs and desktop installers are discontinued.
 - DuckDB is now a mandatory runtime dependency, dynamically linked against the official prebuilt `libduckdb.so` (v1.5.6) with fixed SHA-256 via `scripts/fetch-libduckdb.sh`.
 - Upgraded workspace MSRV to Rust 1.88 (required by `duckdb ~1.10506.1`).
 - Upgraded container runtime and base image to Debian 13 trixie.
 
 ### Fixed
+- Fixed critical false positives in short captures where period candidates exceeded the observed span.
+- Corrected partition counting in `ingest_dir` to take the union of partition dates instead of summing file counts.
+- DuckDB execution errors are now tracked in `files_failed` instead of miscounted as bad rows.
+- Updated `doctor` check status from "skeleton" to "ok" / "degraded".
 - Graceful shutdown on `SIGTERM` when running as PID 1 in container environments.
 
 ### Removed