fix: stop ruflo daemon leak + statusline sqlite3 churn; consolidate shell helpers (#3)

bin/ruflo-enable-learning

@@ -33,19 +33,19 @@ while (( $# )); do
 	shift
 done
 
-if [[ -t 1 ]]; then
-	C_OK=$'\033[32m'; C_WARN=$'\033[33m'; C_FAIL=$'\033[31m'; C_DIM=$'\033[2m'; C_RESET=$'\033[0m'
-else C_OK=""; C_WARN=""; C_FAIL=""; C_DIM=""; C_RESET=""; fi
-ok()   { printf '%s✓%s %s\n' "$C_OK" "$C_RESET" "$*"; }
-warn() { printf '%s⚠%s  %s\n' "$C_WARN" "$C_RESET" "$*"; }
-fail() { printf '%s✗%s %s\n' "$C_FAIL" "$C_RESET" "$*"; }
-dim()  { printf '%s%s%s\n' "$C_DIM" "$*" "$C_RESET"; }
+# Shared helpers (colors, ok/warn/fail/dim, _ruflo_need, ABI/root) from ruflo-lib.sh:
+# installed copy first, then repo-relative (bin/ ↔ ../shell/).
+for _cand in "$HOME/.config/ruflo/ruflo-lib.sh" "$(dirname "${BASH_SOURCE[0]:-$0}")/../shell/ruflo-lib.sh"; do
+	# shellcheck source=/dev/null
+	[ -f "$_cand" ] && { . "$_cand"; break; }
+done
+unset _cand
+[ -n "${_RUFLO_LIB:-}" ] || { echo "ruflo-lib.sh not found — run install.sh" >&2; exit 2; }
 
-command -v node  >/dev/null 2>&1 || { fail "node not on PATH"; exit 2; }
-command -v ruflo >/dev/null 2>&1 || { fail "ruflo not on PATH"; exit 2; }
+_ruflo_need node ruflo
 command -v ruflo-patch-native >/dev/null 2>&1 || { fail "ruflo-patch-native not on PATH (run install.sh)"; exit 2; }
 
-NODE_ABI=$(node -e 'process.stdout.write(process.versions.modules)')
+NODE_ABI=$(_ruflo_node_abi)
 echo "Node ABI $NODE_ABI | ruflo $(ruflo --version 2>/dev/null | tr -d '\n')"
 echo ""
 
@@ -60,7 +60,7 @@ fi
 # On ruflo >= 3.10 the gist's controller-registry patches are already upstream:
 # agentdb resolves >= 3.0 and ReasoningBank gets an embedder. We only WARN if a
 # regression is detected; we do not patch a non-regressed install.
-RUFLO_ROOT="$(npm root -g)/ruflo"
+RUFLO_ROOT="$(_ruflo_global_root)/ruflo"
 MEM="$RUFLO_ROOT/node_modules/@claude-flow/memory"
 ADB_VER=$(node -e "
 try{const p=require.resolve('agentdb',{paths:['$MEM']});process.stdout.write(require(p.split('/agentdb/')[0]+'/agentdb/package.json').version);}catch(e){process.stdout.write('MISSING');}" 2>/dev/null)
@@ -81,9 +81,11 @@ ruvector_repair() {
 		  const {createRequire}=await import('node:module');
 		  const r=createRequire('$d/package.json');
 		  try{ r(r.resolve('$m')); process.exit(0);}catch(e){process.exit(1);}" 2>/dev/null; then
-			( cd "$d" && npm install "$m" --no-save --no-audit --no-fund >/dev/null 2>&1 ) \
-				&& ok "  repaired $m in ${d#$RUFLO_ROOT/node_modules/}" \
-				|| warn "  could not repair $m in ${d#$RUFLO_ROOT/node_modules/}"
+			if ( cd "$d" && npm install "$m" --no-save --no-audit --no-fund >/dev/null 2>&1 ); then
+				ok "  repaired $m in ${d#"$RUFLO_ROOT"/node_modules/}"
+			else
+				warn "  could not repair $m in ${d#"$RUFLO_ROOT"/node_modules/}"
+			fi
 		fi
 	done
 }
@@ -115,24 +117,36 @@ cap() {
 	" 2>/dev/null
 }
 
+# Direct if/then assertions (no eval) — each runs a probe and tallies green/total.
 declare -i green=0 total=0
-report() { total+=1; if eval "$2"; then ok "$1"; green+=1; else fail "$1 — $3"; fi; }
+pass() { total+=1; ok "$1"; green+=1; }
+miss() { total+=1; fail "$1 — $2"; }
 
-report "native better-sqlite3 (no WASM fallback)" \
-	'! echo "$NS" | grep -q "Using sql.js" && echo "$PN" | grep -qE "Nothing to do|already resolve native"' \
-	"still on sql.js/WASM — patch-native did not take"
-report "HNSW vector engine (@ruvector/core → VectorDb)" \
-	'cap "$CLI" "@ruvector/core" "VectorDb"' \
-	"core present but VectorDb missing"
-report "SONA engine (@ruvector/sona → SonaEngine)" \
-	'cap "$NEURAL" "@ruvector/sona" "SonaEngine"' \
-	"sona native module not loadable from @claude-flow/neural"
-report "GNN layer (@ruvector/gnn → RuvectorLayer)" \
-	'cap "$NEURAL" "@ruvector/gnn" "RuvectorLayer"' \
-	"gnn native module not loadable from @claude-flow/neural"
-report "ReasoningBank (agentdb v3)" \
-	'[[ "$ADB_VER" == 3.* ]]' \
-	"agentdb not v3 — ReasoningBank unavailable"
+if ! echo "$NS" | grep -q "Using sql.js" && echo "$PN" | grep -qE "Nothing to do|already resolve native"; then
+	pass "native better-sqlite3 (no WASM fallback)"
+else
+	miss "native better-sqlite3 (no WASM fallback)" "still on sql.js/WASM — patch-native did not take"
+fi
+if cap "$CLI" "@ruvector/core" "VectorDb"; then
+	pass "HNSW vector engine (@ruvector/core → VectorDb)"
+else
+	miss "HNSW vector engine (@ruvector/core → VectorDb)" "core present but VectorDb missing"
+fi
+if cap "$NEURAL" "@ruvector/sona" "SonaEngine"; then
+	pass "SONA engine (@ruvector/sona → SonaEngine)"
+else
+	miss "SONA engine (@ruvector/sona → SonaEngine)" "sona native module not loadable from @claude-flow/neural"
+fi
+if cap "$NEURAL" "@ruvector/gnn" "RuvectorLayer"; then
+	pass "GNN layer (@ruvector/gnn → RuvectorLayer)"
+else
+	miss "GNN layer (@ruvector/gnn → RuvectorLayer)" "gnn native module not loadable from @claude-flow/neural"
+fi
+if [[ "$ADB_VER" == 3.* ]]; then
+	pass "ReasoningBank (agentdb v3)"
+else
+	miss "ReasoningBank (agentdb v3)" "agentdb not v3 — ReasoningBank unavailable"
+fi
 echo ""
 dim "Note: 'ruflo neural status' may still print HNSW/Training as 'Not loaded' — that is"
 dim "a lazy per-process display (getHNSWStatus), not real dormancy. Prove the loop with"

bin/ruflo-learning-verify

@@ -24,17 +24,20 @@ while (( $# )); do
 	esac
 	shift
 done
-if [[ -t 1 ]]; then C_OK=$'\033[32m'; C_FAIL=$'\033[31m'; C_DIM=$'\033[2m'; C_RESET=$'\033[0m'
-else C_OK=""; C_FAIL=""; C_DIM=""; C_RESET=""; fi
-ok()  { printf '%s✓%s %s\n' "$C_OK" "$C_RESET" "$*"; }
-fail(){ printf '%s✗%s %s\n' "$C_FAIL" "$C_RESET" "$*"; }
-dim() { printf '%s%s%s\n' "$C_DIM" "$*" "$C_RESET"; }
+# Shared helpers (colors, ok/fail/dim, _ruflo_need) from ruflo-lib.sh:
+# installed copy first, then repo-relative (bin/ ↔ ../shell/).
+for _cand in "$HOME/.config/ruflo/ruflo-lib.sh" "$(dirname "${BASH_SOURCE[0]:-$0}")/../shell/ruflo-lib.sh"; do
+	# shellcheck source=/dev/null
+	[ -f "$_cand" ] && { . "$_cand"; break; }
+done
+unset _cand
+[ -n "${_RUFLO_LIB:-}" ] || { echo "ruflo-lib.sh not found — run install.sh" >&2; exit 2; }
 
-command -v node  >/dev/null 2>&1 || { fail "node not on PATH"; exit 2; }
-command -v ruflo >/dev/null 2>&1 || { fail "ruflo not on PATH"; exit 2; }
+_ruflo_need node ruflo
 
 T=$(mktemp -d)
 export CLAUDE_FLOW_DB_PATH="$T/.swarm/memory.db"
+# shellcheck disable=SC2329  # invoked via trap below
 cleanup(){ if (( KEEP )); then echo "kept: $T"; else rm -rf "$T"; fi; }
 trap cleanup EXIT
 

bin/ruflo-parity-test

@@ -51,14 +51,16 @@ for arg in "$@"; do
 	esac
 done
 
-# --- terminal helpers -------------------------------------------------------
-if [[ -t 1 ]]; then
-	C_OK=$'\033[32m'; C_FAIL=$'\033[31m'; C_DIM=$'\033[2m'
-	C_HEAD=$'\033[1;36m'; C_RESET=$'\033[0m'
-else
-	C_OK=""; C_FAIL=""; C_DIM=""; C_HEAD=""; C_RESET=""
-fi
+# Shared helpers (colors C_OK/C_FAIL/C_DIM/C_HEAD/C_RESET, _ruflo_daemon_list) from
+# ruflo-lib.sh: installed copy first, then repo-relative (bin/ ↔ ../shell/).
+for _cand in "$HOME/.config/ruflo/ruflo-lib.sh" "$(dirname "${BASH_SOURCE[0]:-$0}")/../shell/ruflo-lib.sh"; do
+	# shellcheck source=/dev/null
+	[ -f "$_cand" ] && { . "$_cand"; break; }
+done
+unset _cand
+[ -n "${_RUFLO_LIB:-}" ] || { echo "ruflo-lib.sh not found — run install.sh" >&2; exit 2; }
 
+# --- test-specific helpers (PASS/FAIL tally, headers) -----------------------
 PASS=0
 FAIL=0
 LOG_FILE=""
@@ -80,8 +82,22 @@ check() {
 	fi
 }
 
+# shellcheck disable=SC2329  # invoked via trap below
 cleanup_on_exit() {
 	local rc=$?
+	# Stop the daemon ruflo-setup-project started for this throwaway workspace
+	# (issue #3) — orphaned whether or not we keep the dir, since the test is done.
+	# The daemon records the RESOLVED (pwd -P) path, so match that too.
+	if [[ -n "${TEST_DIR:-}" ]] && command -v _ruflo_daemon_list >/dev/null 2>&1; then
+		local _rdir; _rdir="$(cd "$TEST_DIR" 2>/dev/null && pwd -P || echo "$TEST_DIR")"
+		local _dp _dws
+		while IFS="$(printf '\t')" read -r _dp _dws; do
+			[[ "$_dws" == "$_rdir" || "$_dws" == "$TEST_DIR" ]] || continue
+			[[ -n "$_dp" ]] && kill "$_dp" 2>/dev/null && log "Stopped test daemon pid=$_dp ($_dws)"
+		done <<EOF
+$(_ruflo_daemon_list)
+EOF
+	fi
 	if (( FAIL == 0 && KEEP == 0 )) && [[ -n "${TEST_DIR:-}" && -d "$TEST_DIR" ]]; then
 		rm -rf "$TEST_DIR"
 		log ""
@@ -91,7 +107,7 @@ cleanup_on_exit() {
 		log "Test dir kept at: $TEST_DIR"
 		[[ -n "$LOG_FILE" ]] && log "Full log:        $LOG_FILE"
 	fi
-	exit $rc
+	exit "$rc"
 }
 trap cleanup_on_exit EXIT INT TERM
 
@@ -143,7 +159,7 @@ LOG_FILE="${TEST_DIR}.log"
 
 head_line "Create test dir"
 mkdir -p "$TEST_DIR"
-cd "$TEST_DIR"
+cd "$TEST_DIR" || { check "test dir created" "cannot cd to $TEST_DIR" "fail"; exit 2; }
 check "test dir created" "$TEST_DIR" "ok"
 
 # Capture full output to a log file alongside the test dir
@@ -188,9 +204,12 @@ if [[ ! -f "$SETTINGS_FILE" ]]; then
 	check "settings.local.json present" "missing" "fail"
 else
 	DB_PATH=$(python3 -c "import json; print(json.load(open('$SETTINGS_FILE')).get('env', {}).get('CLAUDE_FLOW_DB_PATH', ''))" 2>/dev/null)
+	# The literal token Claude Code does NOT expand — matched verbatim, never expanded here.
+	# shellcheck disable=SC2016
+	UNEXPANDED_CPD='${CLAUDE_PROJECT_DIR}'
 	if [[ -z "$DB_PATH" ]]; then
 		check "CLAUDE_FLOW_DB_PATH set" "absent or unreadable" "fail"
-	elif [[ "$DB_PATH" == *'${CLAUDE_PROJECT_DIR}'* ]]; then
+	elif [[ "$DB_PATH" == *"$UNEXPANDED_CPD"* ]]; then
 		check "CLAUDE_FLOW_DB_PATH resolved" "literal \${CLAUDE_PROJECT_DIR} present — Claude Code does NOT expand it; writes will silently fail" "fail"
 	elif [[ "$DB_PATH" != /* ]]; then
 		check "CLAUDE_FLOW_DB_PATH absolute" "not absolute: $DB_PATH" "fail"

bin/ruflo-patch-native

@@ -52,24 +52,20 @@ while (( $# )); do
 	shift
 done
 
-if [[ -t 1 ]]; then
-	C_OK=$'\033[32m'; C_WARN=$'\033[33m'; C_FAIL=$'\033[31m'; C_DIM=$'\033[2m'; C_RESET=$'\033[0m'
-else
-	C_OK=""; C_WARN=""; C_FAIL=""; C_DIM=""; C_RESET=""
-fi
-ok()   { printf '%s✓%s %s\n' "$C_OK" "$C_RESET" "$*"; }
-warn() { printf '%s⚠%s  %s\n' "$C_WARN" "$C_RESET" "$*"; }
-fail() { printf '%s✗%s %s\n' "$C_FAIL" "$C_RESET" "$*"; }
-dim()  { printf '%s%s%s\n' "$C_DIM" "$*" "$C_RESET"; }
+# Shared helpers (colors, ok/warn/fail/dim, _ruflo_need, native bsq3 primitives)
+# from ruflo-lib.sh: installed copy first, then repo-relative (bin/ ↔ ../shell/).
+for _cand in "$HOME/.config/ruflo/ruflo-lib.sh" "$(dirname "${BASH_SOURCE[0]:-$0}")/../shell/ruflo-lib.sh"; do
+	# shellcheck source=/dev/null
+	[ -f "$_cand" ] && { . "$_cand"; break; }
+done
+unset _cand
+[ -n "${_RUFLO_LIB:-}" ] || { echo "ruflo-lib.sh not found — run install.sh" >&2; exit 2; }
 
-command -v node >/dev/null 2>&1 || { fail "node not on PATH"; exit 2; }
-command -v npm  >/dev/null 2>&1 || { fail "npm not on PATH"; exit 2; }
-command -v ruflo >/dev/null 2>&1 || { fail "ruflo not on PATH"; exit 2; }
+_ruflo_need node npm ruflo
 
 NODE_VER=$(node --version)
-NODE_ABI=$(node -e 'process.stdout.write(process.versions.modules)')
-NODE_MAJOR=$(node -e 'process.stdout.write(String(process.versions.node.split(".")[0]))')
-RUFLO_ROOT="$(npm root -g)/ruflo"
+NODE_ABI=$(_ruflo_node_abi)
+RUFLO_ROOT="$(_ruflo_global_root)/ruflo"
 
 echo "Node $NODE_VER (ABI $NODE_ABI) | ruflo $(ruflo --version 2>/dev/null | tr -d '\n')"
 echo "ruflo root: $RUFLO_ROOT"
@@ -107,21 +103,13 @@ if (( ${#AGENTDB_DIRS[@]} == 0 )); then
 fi
 
 needs_patch() {
-	# returns 0 if this agentdb dir cannot resolve a native better-sqlite3
-	local agentdb_dir="$1"
-	node -e "
-try {
-  const p = require.resolve('better-sqlite3', { paths: ['$agentdb_dir'] });
-  const base = p.split('/better-sqlite3/')[0] + '/better-sqlite3';
-  const hasBin = require('fs').existsSync(base + '/build/Release/better_sqlite3.node');
-  process.exit(hasBin ? 1 : 0);   // exit 1 = native present (no patch); 0 = needs patch
-} catch (e) { process.exit(0); }   // unresolvable = needs patch
-" 2>/dev/null
+	# 0 if this agentdb dir cannot resolve a native better-sqlite3 (i.e. needs the patch)
+	! _ruflo_bsq3_is_native "$1"
 }
 
 TO_PATCH=()
 for d in "${AGENTDB_DIRS[@]}"; do
-	rel="${d#$RUFLO_ROOT/node_modules/}"
+	rel="${d#"$RUFLO_ROOT"/node_modules/}"
 	if needs_patch "$d"; then
 		echo "  needs patch: $rel"
 		TO_PATCH+=("$d")
@@ -146,7 +134,7 @@ echo "Installing better-sqlite3@$BSQ_RANGE into ${#TO_PATCH[@]} location(s)..."
 PATCH_LOG=$(mktemp)
 patched=0
 for d in "${TO_PATCH[@]}"; do
-	rel="${d#$RUFLO_ROOT/node_modules/}"
+	rel="${d#"$RUFLO_ROOT"/node_modules/}"
 	if ( cd "$d" && npm install "better-sqlite3@$BSQ_RANGE" --no-save --no-audit --no-fund >>"$PATCH_LOG" 2>&1 ); then
 		ok "patched $rel"
 		patched=$((patched+1))
@@ -160,7 +148,7 @@ echo ""
 echo "Verifying native binaries..."
 allnative=1
 for d in "${TO_PATCH[@]}"; do
-	rel="${d#$RUFLO_ROOT/node_modules/}"
+	rel="${d#"$RUFLO_ROOT"/node_modules/}"
 	if needs_patch "$d"; then
 		fail "$rel still not native"
 		allnative=0
@@ -179,7 +167,7 @@ echo ""
 # --- Functional smoke test --------------------------------------------------
 echo "Smoke test (store → retrieve via native path)..."
 TMPD=$(mktemp -d)
-( cd "$TMPD"
+( cd "$TMPD" || exit 1
   export CLAUDE_FLOW_DB_PATH="$TMPD/.swarm/memory.db"
   ruflo init --minimal --force >/dev/null 2>&1
   ruflo memory init >/dev/null 2>&1

bin/ruflo-security-verify

@@ -24,15 +24,17 @@ while (( $# )); do
 	esac
 	shift
 done
-if [[ -t 1 ]]; then C_OK=$'\033[32m'; C_WARN=$'\033[33m'; C_FAIL=$'\033[31m'; C_RESET=$'\033[0m'
-else C_OK=""; C_WARN=""; C_FAIL=""; C_RESET=""; fi
-ok()  { printf '%s✓%s %s\n' "$C_OK" "$C_RESET" "$*"; }
-warn(){ printf '%s⚠%s  %s\n' "$C_WARN" "$C_RESET" "$*"; }
-fail(){ printf '%s✗%s %s\n' "$C_FAIL" "$C_RESET" "$*"; }
+# Shared helpers (colors, ok/warn/fail, _ruflo_need, global-root) from ruflo-lib.sh:
+# installed copy first, then repo-relative (bin/ ↔ ../shell/).
+for _cand in "$HOME/.config/ruflo/ruflo-lib.sh" "$(dirname "${BASH_SOURCE[0]:-$0}")/../shell/ruflo-lib.sh"; do
+	# shellcheck source=/dev/null
+	[ -f "$_cand" ] && { . "$_cand"; break; }
+done
+unset _cand
+[ -n "${_RUFLO_LIB:-}" ] || { echo "ruflo-lib.sh not found — run install.sh" >&2; exit 2; }
 
-command -v node  >/dev/null 2>&1 || { fail "node not on PATH"; exit 2; }
-command -v ruflo >/dev/null 2>&1 || { fail "ruflo not on PATH"; exit 2; }
-RUFLO_ROOT="$(npm root -g)/ruflo"
+_ruflo_need node ruflo
+RUFLO_ROOT="$(_ruflo_global_root)/ruflo"
 declare -i bad=0
 
 # 1. modules load
@@ -71,4 +73,8 @@ if ruflo security cve --list 2>&1 | grep -qi "no cve database"; then
 fi
 
 echo ""
-(( bad == 0 )) && { ok "Security surface verified."; exit 0; } || { fail "$bad security capability/ies need attention."; exit 1; }
+if (( bad == 0 )); then
+	ok "Security surface verified."; exit 0
+else
+	fail "$bad security capability/ies need attention."; exit 1
+fi

docs/TROUBLESHOOTING.md

@@ -179,3 +179,43 @@ ruflo cleanup --force               # remove ruflo artifacts (dry-run by default
 rm -rf .swarm .claude-flow .mcp.json
 ruflo-setup-project                 # re-create cleanly
 ```
+
+## Claude Code crashes with ENOSPC / orphan `ruflo daemon` processes pile up
+
+**Symptom.** Claude Code dies mid-session with
+`the temp filesystem at /private/tmp/claude-501/<project>/<uuid>/tasks is full
+(0MB free)`, and/or `ps axww | grep "daemon start"` shows many `ruflo daemon`
+processes — some pointed at `--workspace` directories that no longer exist
+(e.g. `/tmp/test-*` from `ruflo-parity-test` runs). See issue #3.
+
+**Why.** `ruflo-setup-project` starts a per-workspace `ruflo daemon` (this is what
+makes self-learning continuous). Before this fix, nothing stopped them, so each
+throwaway/removed workspace left a daemon running forever. Separately, the
+statusline footer used to spawn several `sqlite3` subprocesses on every render;
+that volume of captured subprocess output is what fills Claude Code's size-limited
+sandbox `tasks` tmpfs.
+
+**Fix is built in now.** The statusline footer caches its QE metrics
+(`RUFLO_QE_STATUSLINE_TTL_MS`, default 60000ms) and makes at most one `sqlite3`
+call per TTL window. The daemon start is idempotent per workspace, and orphans are
+reaped.
+
+**Reap existing orphans:**
+
+```bash
+ruflo-daemon-gc            # list daemons whose --workspace is gone
+ruflo-daemon-gc --kill     # stop exactly those (live-project daemons untouched)
+```
+
+`uninstall.sh` also reaps stale daemons; `uninstall.sh --this-project` additionally
+stops the current repo's daemon.
+
+**If you run many hooks and still hit the tmpfs limit**, point Claude Code's
+subprocess tmpdir at your main filesystem (more space than the sandbox tmpfs) by
+adding to `~/.claude/settings.json`:
+
+```json
+{ "env": { "CLAUDE_CODE_TMPDIR": "/Users/<you>/tmp/claude-code" } }
+```
+
+Create the directory first (`mkdir -p ~/tmp/claude-code`).