feat(discord-bridge): punte Discord -> Claude Code pe LXC 171

Implementeaza planul claude-master-plan-discord-bridge-20260830 (15 taskuri,
3 lane-uri paralele) — un bot subtire discord.py peste CLI-ul `claude`, cu
proces persistent per fir alimentat pe stdin cu --input-format stream-json.

Nucleu: runner (proces persistent + reaper 20min + respawn --resume), stream
(parser tolerant), session_store (scriere atomica, lock per fir, detectare PID
reuse, recovery), limits (max 4 procese, timeout tur, rate per user, plafon cost
pe zi), render (un loop de editare per canal, interval adaptiv).

Adaptor: allowlist guild/canal/user fail-closed cu respingerea webhook-urilor,
comenzi !new/!cd/!model/!status/!stop/!cleanup, cost si model in subsolul
fiecarui raspuns. Mesajul sosit in timpul unui tur devine steering, nu tur nou.

Securitate: hook PreToolUse fail-closed care cere confirmare in Discord pentru
operatiuni ireversibile, wrapper `infra` cu lista explicita de hosturi. Deny
rules raman strat cosmetic, nu bariera (verificat: /usr/bin/ssh trece pe langa).

Ops: alerte email pe conventia repo-ului, !cleanup pentru orfani, unit systemd
user cu KillMode=control-group si limite de memorie, install.sh idempotent.

Verificat: 275 teste fara retea/Discord/API (10.8s), identic cu si fara
discord.py instalat; e2e pe CLI real confirma steering-ul mid-tur (mesaj la 6s
intr-un tool call de 25s schimba raspunsul final).

Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01B29CApsP1JkSdjYaGaHpE7

proxmox/lxc171-claude-agent/discord-bridge/ops/claude-discord.service

@@ -0,0 +1,61 @@
+# Unit systemd de UTILIZATOR pentru puntea Discord -> Claude Code (LXC 171).
+#
+# Se instaleaza in ~/.config/systemd/user/claude-discord.service (vezi install.sh).
+# Rulare ca utilizatorul `claude`, NU ca root: puntea are exact accesul pe care il
+# are omul la terminal — decizie de arhitectura, nu scapare (vezi README, Securitate).
+#
+#   systemctl --user daemon-reload
+#   systemctl --user enable --now claude-discord
+#   systemctl --user status claude-discord
+#   journalctl --user -u claude-discord -f
+
+[Unit]
+Description=Punte Discord -> Claude Code (LXC 171 claude-agent)
+Documentation=file:///workspace/romfastsql/proxmox/lxc171-claude-agent/discord-bridge/README.md
+After=network-online.target
+Wants=network-online.target
+
+# Crash loop: dupa 5 porniri esuate in 300s systemd renunta si lasa unitul in
+# `failed`, in loc sa se invarta la nesfarsit. Botul trimite alerta pe email
+# inainte sa moara (alerts.alert_crash_loop).
+StartLimitIntervalSec=300
+StartLimitBurst=5
+
+[Service]
+Type=simple
+WorkingDirectory=%h/.claude-discord
+
+# Token-ul Discord si allowlist-ul stau in fisier 0600, nu in unit.
+EnvironmentFile=%h/.claude-discord/env
+Environment=PYTHONUNBUFFERED=1
+# CLI-ul `claude` e instalat prin nvm; PATH-ul unui unit de utilizator nu-l contine.
+Environment=PATH=%h/.nvm/versions/node/v20.19.6/bin:/usr/local/sbin:/usr/local/bin:/usr/sbin:/usr/bin:/sbin:/bin
+
+ExecStartPre=/bin/mkdir -p %h/.claude-discord/logs
+ExecStart=%h/.claude-discord/venv/bin/python /workspace/romfastsql/proxmox/lxc171-claude-agent/discord-bridge/bot.py
+
+# ESENTIAL: la stop/restart omoara TOT cgroup-ul, nu doar procesul principal.
+# Fara asta procesele `claude` (406 MB RSS masurat) raman in urma si umplu
+# containerul. Nu prinde totusi ce s-a desprins din cgroup — pentru aia exista
+# comanda `!cleanup` (cleanup.py).
+KillMode=control-group
+KillSignal=SIGTERM
+TimeoutStopSec=30
+
+Restart=on-failure
+RestartSec=5
+# Backoff exponential: 5s, 10s, 20s, ... plafonat la 120s (systemd >= 254).
+RestartSteps=5
+RestartMaxDelaySec=120
+
+# Containerul are istoric de OOM (incident 2026-06-24). MemoryHigh strange
+# procesul prin reclaim inainte sa se ajunga la OOM-kill la MemoryMax.
+MemoryHigh=3G
+MemoryMax=6G
+
+StandardOutput=append:%h/.claude-discord/logs/bot.log
+StandardError=append:%h/.claude-discord/logs/bot.log
+SyslogIdentifier=claude-discord
+
+[Install]
+WantedBy=default.target