fix: cut log noise and add severity-based searching

README.md

@@ -98,10 +98,20 @@ agent/Executor filesystem boundary:
 
 ```bash
 tx9 logs media-bot --source executor,codex --since 24h --grep failed
+tx9 logs media-bot --level warn --since 24h
 tx9 logs media-bot --json | jq .
-tx9 logs export media-bot --output media-bot-logs.tar.gz
+tx9 logs export media-bot --level warn --output media-bot-logs.tar.gz
 ```
 
+Events carry a best-effort `level` (debug/info/warn/error) parsed from known
+log formats; `--level` keeps that level and above, counting unleveled events
+as info. The same threshold filters an export's `events.jsonl` and is recorded
+as `manifest.filters.level`. Steady-state noise is kept out of the durable
+streams: the reconcile loop logs Executor reachability only when it changes,
+and capture collapses consecutive identical lines into one event plus a
+repeat-count summary while raw text logs stay byte-faithful. Multi-line Hermes
+records (tracebacks, embedded files) are grouped into single events.
+
 Known bearer-token forms are redacted by default. Log exports are created
 mode `0600`; they can still contain prompts, tool output, and private work, so
 treat them as sensitive. Existing boxes begin collecting the new durable

docs/tx9-cli-design.md

@@ -52,7 +52,7 @@ Docker overview can be read. `tx9 help` remains Docker-independent.
 | `tx9 backup <box>` (aliases `export`, `save`) | Flags: `--path` (default `~/Downloads`), `--password`/env/prompt, `--no-encrypt`. Quiesce → archive agent /data → validate → (encrypt) → verify → `<box>-<timestamp>.tx9`. |
 | `tx9 import <file.tx9>` (aliases `load`, `restore`) | Flags: `--name`, `--password`/env/prompt. Validate before creating anything; restore staged; arrive quiesced + gateway-disabled + fresh token; fail on name collision. |
 | `tx9 mount <add\|list\|remove> ...` | Persist host-directory bind mounts for an agent and recreate only its disposable container. Targets must be below `/mnt`, outside the portable `/data` volume. `add` supports `--read-only` and `--require-mountpoint`. |
-| `tx9 logs <box>` | Query durable agent, Executor, Hermes, Codex, and Claude events. Filters include source, age, text, count, and normalized JSONL. `tx9 logs export <box>` creates a mode-0600 portable log bundle from both isolated volumes. |
+| `tx9 logs <box>` | Query durable agent, Executor, Hermes, Codex, and Claude events. Filters include source, age, text, severity (`--level`, this level and above; unleveled events count as info), count, and normalized JSONL. `tx9 logs export <box>` creates a mode-0600 portable log bundle from both isolated volumes. |
 | `tx9 resources <box>` | Show actual container CPU/RAM limits and volume usage versus advisory budgets. `resources set` updates limits live and persists them; `resources reset` restores 4 CPU/8 GiB (agent) and 2 CPU/2 GiB (Executor). |
 | `tx9 gateway <status\|enable\|disable> <box>` | Inspect or control the container-supervised Hermes gateway. Enable requires `--confirm-single-writer`. |
 | `tx9 open <box>` | Print (or open) the authenticated dashboard URL (`?_token=`). |

guest/hb

@@ -41,6 +41,25 @@ EXECUTOR_HOST="${EXECUTOR_HOST:-127.0.0.1}"
 _exec_remote() { [[ "$EXECUTOR_HOST" != 127.0.0.1 ]]; }
 _executor_url() { printf 'http://%s:%s/mcp' "$EXECUTOR_HOST" "${EXECUTOR_PORT:-4788}"; }
 
+# Steady-state reporting: the reconcile loop re-checks Executor every cycle,
+# so unchanged results must not re-log — they flooded the durable agent
+# stream with one identical line per cycle. Non-durable state under /tmp
+# resets with the container, so the first check after a restart always logs.
+HB_RUNTIME_STATE_DIR="${HB_RUNTIME_STATE_DIR:-/tmp/hermes-box-runtime}"
+HB_STEADY_QUIET="${HB_STEADY_QUIET:-0}"
+_steady_report() {  # <key> <state> <message> [stderr]
+  local key="$1" state="$2" message="$3" fd="${4:-stdout}" state_file
+  if [[ "$HB_STEADY_QUIET" != 1 ]]; then
+    if [[ "$fd" == stderr ]]; then echo "$message" >&2; else echo "$message"; fi
+    return 0
+  fi
+  state_file="$HB_RUNTIME_STATE_DIR/$key.state"
+  [[ "$(cat "$state_file" 2>/dev/null)" != "$state" ]] || return 0
+  mkdir -p "$HB_RUNTIME_STATE_DIR" 2>/dev/null || true
+  printf '%s' "$state" >"$state_file" 2>/dev/null || true
+  if [[ "$fd" == stderr ]]; then echo "$message" >&2; else echo "$message"; fi
+}
+
 init() {
   local managed=(
     "$AGENT_HOME" "$AGENT_HOME/.claude" "$AGENT_HOME/.codex" "$AGENT_HOME/.hermes"
@@ -152,8 +171,8 @@ _spawn_without_lock_fd() {
 _executor_up() {
   init
   if _exec_remote; then
-    _wait_port open || { echo "remote executor unreachable at $EXECUTOR_HOST:${EXECUTOR_PORT:-4788}" >&2; return 1; }
-    echo "executor daemon: reachable (remote)"
+    _wait_port open || { _steady_report executor unreachable "remote executor unreachable at $EXECUTOR_HOST:${EXECUTOR_PORT:-4788}" stderr; return 1; }
+    _steady_report executor reachable "executor daemon: reachable (remote)"
     return 0
   fi
   command -v executor >/dev/null 2>&1 || {
@@ -163,7 +182,7 @@ _executor_up() {
   }
   _acquire_daemon_lock "$EXECUTOR_LOCK" || { echo "could not acquire Executor lock" >&2; return 1; }
   if _port_open; then
-    echo "executor daemon: running"
+    _steady_report executor running "executor daemon: running"
     _release_daemon_lock "$EXECUTOR_LOCK"
     return 0
   fi
@@ -323,7 +342,7 @@ up() {
 reconcile() {
   init
   [[ ! -e "$QUIESCE_FILE" ]] || return 0
-  _executor_up || return 1
+  HB_STEADY_QUIET=1 _executor_up || return 1
   _start_gateway
 }