feat(orchestrator): board-roll.sh — auto-roll LIVE board to LEDGER under byte cap (#868)

feat(orchestrator): board-roll.sh - auto-roll LIVE board to LEDGER under byte cap

Closes #868

packages/mosaic/framework/tools/orchestrator/README.md

@@ -0,0 +1,85 @@
+# orchestrator/ tools
+
+Helper scripts for r0 coordinator / orchestrator sessions — mission lifecycle,
+session health, continuation, and board maintenance. See
+`framework/guides/ORCHESTRATOR-PROTOCOL.md` for the surrounding process.
+
+| Script | Purpose |
+|--------|---------|
+| `mission-init.sh` | Initialize a new orchestration mission (manifest, scratchpad, TASKS.md). |
+| `mission-status.sh` | Show the mission progress dashboard. |
+| `session-run.sh` | Generate continuation context and launch the target runtime. |
+| `session-resume.sh` | Crash recovery for dead orchestrator sessions. |
+| `session-status.sh` | Check agent session health. |
+| `continue-prompt.sh` | Generate the continuation prompt for the next session. |
+| `board-roll.sh` | Keep a LIVE orchestration board under its byte cap by rolling the oldest entries to its LEDGER. |
+| `smoke-test.sh` | Behavior smoke checks for the coord continue/run workflows. |
+| `test-board-roll.sh` | Regression harness for `board-roll.sh`. |
+| `_lib.sh` | Shared functions sourced by the above (state files, TASKS.md parsing, locks). |
+
+## board-roll.sh
+
+Coordinator boards (`MOS-ORCHESTRATION-BOARD-LIVE.md`, `MS-LEAD-BOARD-LIVE.md`)
+follow a **"< 8 KB LIVE"** discipline: the LIVE board is the only file loaded on
+resume, so it must stay small, and history lives in an append-only LEDGER. When a
+board write would push LIVE over its cap, coordinators otherwise hand-trim and
+retry every time — an observed 38 ABORT-OVER-CAP cycles in one 24 h window.
+`board-roll.sh` automates that trim mechanically and reversibly: the audit trail
+is moved to the LEDGER instead of being hand-deleted.
+
+### Contract (conservative — it never guesses what is safe to move)
+
+The LIVE board opts in by wrapping its aging archival ticks in an explicit roll
+zone. Everything **outside** the markers (title, protocol blockquote, curated
+always-current `##` sections) is pinned and never touched:
+
+```markdown
+# MOS ORCHESTRATION BOARD — LIVE state
+> protocol blockquote … (pinned)
+
+## 🟦 Curated always-current section (pinned)
+…
+
+<!-- BOARD-ROLL:START -->
+### 2026-07-22 (mid²²) — newest tick, stays longest
+…
+### 2026-07-20 (dawn) — oldest tick, rolled first
+…
+<!-- BOARD-ROLL:END -->
+```
+
+Inside the zone, entries are delimited by a heading marker (default `### `) and
+are assumed **newest-first (top) → oldest-last (bottom)**. `board-roll.sh` moves
+whole oldest (bottom-most) entry blocks out of the zone and appends them verbatim
+to the LEDGER, one at a time, until LIVE is back under the cap or the zone is
+empty. If the board has no markers, it exits `3` and changes nothing — adding the
+markers is a deliberate opt-in by the board owner.
+
+### Usage
+
+```bash
+board-roll.sh --live <LIVE.md> --ledger <LEDGER.md> [options]
+
+  --live   <path>    LIVE board file (required)
+  --ledger <path>    append-only LEDGER file (required; created if absent)
+  --cap    <bytes>   size ceiling for LIVE (default 8192)
+  --marker <prefix>  entry-heading prefix inside the roll zone (default "### ")
+  --dry-run          report what would move; change nothing
+  -h, --help         show help and exit 0
+```
+
+Only **one** roll zone is supported. If a board carries more than one
+`BOARD-ROLL:START`/`END` pair, `board-roll.sh` refuses (exit `3`, zero changes)
+rather than span first-START..last-END and relocate the curated content between
+the zones — consolidate the ticks into a single zone instead.
+
+Exit codes: `0` LIVE under cap (already, or after rolling) — on `--dry-run`, a
+plan exists or nothing to do · `2` usage / argument / IO error · `3` cannot meet
+the cap (no markers, **more than one marker pair**, or the pinned sections alone
+exceed the cap and need a manual trim).
+
+Writes are atomic (temp file + `mv`, LEDGER first) so a failure never leaves a
+board half-written; line endings are normalized to LF on rewrite. `--dry-run`
+first is recommended when wiring it into a board update protocol.
+
+Run the regression suite with `bash test-board-roll.sh`.

packages/mosaic/framework/tools/orchestrator/board-roll.sh

@@ -0,0 +1,277 @@
+#!/usr/bin/env bash
+#
+# board-roll.sh — keep a LIVE orchestration board under its byte cap by rolling
+#                 the oldest archival entries out to its append-only LEDGER.
+#
+# WHY: coordinator boards (MOS-ORCHESTRATION-BOARD-LIVE.md, MS-LEAD-BOARD-LIVE.md)
+# enforce a "< 8 KB LIVE" discipline via a self-guard that ABORTs the board write
+# when the file exceeds the cap. In practice the LIVE board keeps bumping the cap,
+# so coordinators hand-trim + retry every time (observed: 38 ABORT-OVER-CAP cycles
+# in a 24h window on one coordinator). This automates that trim, mechanically and
+# reversibly, so the audit trail is preserved in the LEDGER instead of hand-deleted.
+#
+# CONTRACT (conservative by design — it NEVER guesses what is safe to move):
+#   The LIVE board must declare an explicit ROLL ZONE with HTML-comment markers:
+#
+#       <!-- BOARD-ROLL:START -->
+#       ### 2026-07-22 (newest tick — stays longest)
+#       ...
+#       ### 2026-07-19 (oldest tick — rolled first)
+#       ...
+#       <!-- BOARD-ROLL:END -->
+#
+#   Everything OUTSIDE the markers (title, protocol blockquote, curated always-current
+#   `##` sections) is PINNED and never touched. Inside the zone, entries are delimited
+#   by a heading marker (default `### `) and are assumed newest-first (top) → oldest-last
+#   (bottom), matching board convention. board-roll moves whole oldest (bottom-most)
+#   entry blocks out of the zone and APPENDS them verbatim to the LEDGER, one block at a
+#   time, until the LIVE file is back under the cap or the zone is empty.
+#
+#   If no markers are present, it exits 3 without changing anything (safe default —
+#   adding the markers is a deliberate opt-in by the board owner).
+#
+# USAGE:
+#   board-roll.sh --live <LIVE.md> --ledger <LEDGER.md> [options]
+#
+# OPTIONS:
+#   --live   <path>   LIVE board file (required)
+#   --ledger <path>   append-only LEDGER file (required; created if absent)
+#   --cap    <bytes>  size ceiling for LIVE (default 8192)
+#   --marker <prefix> entry-heading prefix inside the roll zone (default "### ")
+#   --dry-run         report what would move + resulting size; change nothing
+#   -h, --help        print usage and exit 0
+#
+# EXIT CODES:
+#   0  LIVE is under cap (already, or after rolling); on --dry-run, 0 = a plan exists
+#      (or nothing to do)
+#   2  usage / argument / IO error (bad flag, missing file, unwritable target)
+#   3  cannot satisfy the cap: no roll markers present, MORE THAN ONE marker pair
+#      (multiple zones are refused, not guessed), OR the zone was emptied and LIVE
+#      is still over cap (curated pinned sections need a manual trim)
+#
+# NOTE: line endings are normalized to LF on rewrite (boards are LF markdown); a
+# trailing newline is always ensured. Writes are atomic (temp file + mv) so a
+# failure never leaves LIVE or LEDGER half-written.
+
+set -euo pipefail
+
+START_MARK='<!-- BOARD-ROLL:START -->'
+END_MARK='<!-- BOARD-ROLL:END -->'
+
+usage() {
+    cat <<'EOF'
+Usage: board-roll.sh --live <LIVE.md> --ledger <LEDGER.md> [options]
+
+Roll the oldest entries out of a LIVE orchestration board into its LEDGER
+until the LIVE file is under a byte cap. Conservative: only content inside
+explicit <!-- BOARD-ROLL:START -->/<!-- BOARD-ROLL:END --> markers is moved.
+
+Options:
+  --live   <path>    LIVE board file (required)
+  --ledger <path>    append-only LEDGER file (required; created if absent)
+  --cap    <bytes>   size ceiling for LIVE (default 8192)
+  --marker <prefix>  entry-heading prefix inside the roll zone (default "### ")
+  --dry-run          report what would move; change nothing
+  -h, --help         show this help and exit 0
+
+Exit: 0 under cap (or dry-run plan) · 2 usage/IO error · 3 cannot meet cap
+      (no markers, or pinned sections alone exceed the cap).
+EOF
+}
+
+die() { echo "board-roll: $*" >&2; exit 2; }
+
+LIVE=""; LEDGER=""; CAP=8192; MARKER='### '; DRYRUN=0
+while [[ $# -gt 0 ]]; do
+    case "$1" in
+        --live)    LIVE="${2:-}"; shift 2 || die "--live needs a value" ;;
+        --ledger)  LEDGER="${2:-}"; shift 2 || die "--ledger needs a value" ;;
+        --cap)     CAP="${2:-}"; shift 2 || die "--cap needs a value" ;;
+        --marker)  MARKER="${2:-}"; shift 2 || die "--marker needs a value" ;;
+        --dry-run) DRYRUN=1; shift ;;
+        -h|--help) usage; exit 0 ;;
+        *) usage >&2; die "unknown option: $1" ;;
+    esac
+done
+
+[[ -n "$LIVE"   ]] || { usage >&2; die "--live is required"; }
+[[ -n "$LEDGER" ]] || { usage >&2; die "--ledger is required"; }
+[[ -f "$LIVE"   ]] || die "LIVE file not found: $LIVE"
+[[ "$CAP" =~ ^[0-9]+$ ]] || die "--cap must be a non-negative integer, got: $CAP"
+
+# --- read LIVE into a line array (newlines stripped; re-added on write) ---------
+mapfile -t LINES < "$LIVE"
+
+# byte size of an array rendered as LF-terminated text
+render_size() {
+    if [[ $# -eq 0 ]]; then printf 0; return; fi
+    printf '%s\n' "$@" | wc -c
+}
+
+orig_size=$(render_size "${LINES[@]}")
+
+# --- already under cap → nothing to do -----------------------------------------
+if (( orig_size < CAP )); then
+    echo "board-roll: LIVE is ${orig_size}B (< cap ${CAP}B) — nothing to roll."
+    exit 0
+fi
+
+# --- locate the roll-zone markers ----------------------------------------------
+# Exactly ONE marker pair is supported. If a board carries more than one START or
+# END marker we REFUSE (exit 3, zero changes) rather than guess: a naive
+# first-START..last-END span would swallow the curated content and the intermediate
+# markers sitting between two intended zones and silently relocate that pinned text
+# to the LEDGER — the exact data-loss this tool exists to prevent. Refusing matches
+# the "no markers = exit 3" conservative posture.
+start_idx=-1; end_idx=-1; start_count=0; end_count=0
+for i in "${!LINES[@]}"; do
+    if [[ "${LINES[$i]}" == "$START_MARK" ]]; then
+        if (( start_count == 0 )); then start_idx=$i; fi
+        start_count=$(( start_count + 1 ))
+    fi
+    if [[ "${LINES[$i]}" == "$END_MARK" ]]; then
+        end_idx=$i
+        end_count=$(( end_count + 1 ))
+    fi
+done
+if (( start_count > 1 || end_count > 1 )); then
+    echo "board-roll: LIVE is ${orig_size}B (>= cap ${CAP}B) but has ${start_count} START / ${end_count} END" >&2
+    echo "  markers — only a SINGLE '$START_MARK' … '$END_MARK' roll zone is supported." >&2
+    echo "  Multiple zones are refused (not guessed) so content between zones is never relocated." >&2
+    echo "  Consolidate the archival ticks into one zone, or trim manually." >&2
+    exit 3
+fi
+if (( start_idx < 0 || end_idx < 0 || end_idx <= start_idx )); then
+    echo "board-roll: LIVE is ${orig_size}B (>= cap ${CAP}B) but no usable roll zone" >&2
+    echo "  (need '$START_MARK' then '$END_MARK'). Add the markers around the" >&2
+    echo "  archival tick section to opt this board into automatic rolling." >&2
+    exit 3
+fi
+
+# preamble = lines [0 .. start_idx]   (inclusive of START marker)
+# zone     = lines (start_idx .. end_idx)  (exclusive of both markers)
+# footer   = lines [end_idx .. end]   (inclusive of END marker)
+preamble=(); zone=(); footer=()
+for i in "${!LINES[@]}"; do
+    if   (( i <= start_idx )); then preamble+=("${LINES[$i]}")
+    elif (( i <  end_idx  )); then zone+=("${LINES[$i]}")
+    else footer+=("${LINES[$i]}")
+    fi
+done
+
+# --- split the zone into a fixed head + entry blocks ----------------------------
+# zone_head = any zone lines before the first entry marker (kept, never rolled).
+# blocks[k] = newline-joined text of entry k (marker line .. line before next marker).
+zone_head=(); declare -a block_start=()
+first_block=-1
+for i in "${!zone[@]}"; do
+    if [[ "${zone[$i]}" == "$MARKER"* ]]; then
+        [[ $first_block -eq -1 ]] && first_block=$i
+        block_start+=("$i")
+    fi
+done
+if (( first_block == -1 )); then
+    echo "board-roll: LIVE is ${orig_size}B (>= cap ${CAP}B) but the roll zone has no" >&2
+    echo "  '${MARKER}' entries to move. Trim the pinned sections manually." >&2
+    exit 3
+fi
+for (( i=0; i<first_block; i++ )); do zone_head+=("${zone[$i]}"); done
+
+nblocks=${#block_start[@]}
+# block k spans zone[ block_start[k] .. (block_start[k+1]-1 or end-of-zone) ]
+block_text() {  # $1 = block index → prints the block's lines, LF-joined (no trailing)
+    local k=$1 s e
+    s=${block_start[$k]}
+    if (( k+1 < nblocks )); then e=$(( block_start[$((k+1))] - 1 )); else e=$(( ${#zone[@]} - 1 )); fi
+    local out=()
+    for (( j=s; j<=e; j++ )); do out+=("${zone[$j]}"); done
+    printf '%s\n' "${out[@]}"
+}
+
+# --- greedily roll oldest (bottom-most) blocks until under cap ------------------
+# keep = number of newest blocks retained; start with all, drop from the bottom.
+keep=$nblocks       # blocks [keep .. nblocks-1] are the oldest set that gets moved
+current_size=$orig_size
+build_live_size() {  # size of LIVE if we keep blocks [0 .. keep-1]
+    local acc=("${preamble[@]}" "${zone_head[@]}")
+    local k s e j
+    for (( k=0; k<keep; k++ )); do
+        s=${block_start[$k]}
+        if (( k+1 < nblocks )); then e=$(( block_start[$((k+1))] - 1 )); else e=$(( ${#zone[@]} - 1 )); fi
+        for (( j=s; j<=e; j++ )); do acc+=("${zone[$j]}"); done
+    done
+    acc+=("${footer[@]}")
+    render_size "${acc[@]}"
+}
+while (( current_size >= CAP && keep > 0 )); do
+    keep=$(( keep - 1 ))
+    current_size=$(build_live_size)
+done
+
+moved_count=$(( nblocks - keep ))
+if (( moved_count == 0 )); then
+    # zone had entries but none movable brought us under (shouldn't happen: keep hits 0)
+    echo "board-roll: could not reduce LIVE below cap (${current_size}B >= ${CAP}B)." >&2
+    exit 3
+fi
+
+# --- dry-run report -------------------------------------------------------------
+plan_headers() {
+    local k
+    for (( k=keep; k<nblocks; k++ )); do
+        # first line of each moved block
+        printf '    %s\n' "${zone[${block_start[$k]}]}"
+    done
+}
+if (( DRYRUN )); then
+    echo "board-roll: DRY RUN"
+    echo "  LIVE now:   ${orig_size}B (cap ${CAP}B) — over by $(( orig_size - CAP ))B"
+    echo "  would roll: ${moved_count} of ${nblocks} entr$([[ $moved_count -eq 1 ]] && echo y || echo ies) (oldest first):"
+    plan_headers
+    echo "  LIVE after: ${current_size}B"
+    if (( current_size >= CAP )); then
+        echo "  WARNING: still >= cap after emptying the zone; pinned sections need a manual trim." >&2
+        exit 3
+    fi
+    exit 0
+fi
+
+# --- commit the roll atomically -------------------------------------------------
+live_tmp="$(mktemp "${LIVE}.roll.XXXXXX")" || die "cannot create temp next to LIVE"
+ledger_tmp=""
+# shellcheck disable=SC2329  # invoked indirectly via `trap cleanup EXIT`
+cleanup() { rm -f "$live_tmp" "$ledger_tmp" 2>/dev/null || true; }
+trap cleanup EXIT
+
+# new LIVE = preamble + zone_head + kept blocks + footer
+{
+    printf '%s\n' "${preamble[@]}" "${zone_head[@]}"
+    for (( k=0; k<keep; k++ )); do block_text "$k"; done
+    printf '%s\n' "${footer[@]}"
+} > "$live_tmp"
+
+# LEDGER gets the moved blocks appended verbatim, in original top→bottom order,
+# under a provenance separator. LEDGER is append-only, so we only ever add at EOF.
+ledger_tmp="$(mktemp "${LEDGER}.roll.XXXXXX")" || die "cannot create temp next to LEDGER"
+if [[ -f "$LEDGER" ]]; then cat "$LEDGER" > "$ledger_tmp"; fi
+# ensure a trailing newline on existing content before appending
+if [[ -s "$ledger_tmp" && -n "$(tail -c1 "$ledger_tmp")" ]]; then printf '\n' >> "$ledger_tmp"; fi
+{
+    printf '\n<!-- board-roll: %d entr%s rolled from %s -->\n' \
+        "$moved_count" "$([[ $moved_count -eq 1 ]] && echo y || echo ies)" "$(basename "$LIVE")"
+    for (( k=keep; k<nblocks; k++ )); do block_text "$k"; done
+} >> "$ledger_tmp"
+
+# atomic swap (both, LEDGER first so a crash never drops content that left LIVE)
+mv "$ledger_tmp" "$LEDGER"; ledger_tmp=""
+mv "$live_tmp" "$LIVE"; live_tmp=""
+trap - EXIT
+
+# read the real on-disk size back (truthful, not the predicted value)
+final_size=$(wc -c < "$LIVE")
+echo "board-roll: rolled ${moved_count} entr$([[ $moved_count -eq 1 ]] && echo y || echo ies) to $(basename "$LEDGER"); LIVE ${orig_size}B → ${final_size}B (cap ${CAP}B)."
+if (( final_size >= CAP )); then
+    echo "board-roll: still >= cap after rolling all zone entries; pinned sections need a manual trim." >&2
+    exit 3
+fi
+exit 0

packages/mosaic/framework/tools/orchestrator/test-board-roll.sh

@@ -0,0 +1,156 @@
+#!/usr/bin/env bash
+# Regression harness for board-roll.sh — rolling oldest LIVE-board entries to LEDGER.
+#
+# Asserts:
+#   1. Under cap → no-op, exit 0, files unchanged.
+#   2. Over cap, no roll markers → exit 3, LIVE unchanged (never guesses).
+#   3. Over cap, markers present → rolls the fewest oldest entries to get under cap,
+#      LIVE ends under cap, pinned preamble/footer + newest entries preserved.
+#   4. Rolled blocks land in the LEDGER verbatim, oldest set in original order.
+#   5. --dry-run changes nothing and reports a plan.
+#   6. Zone emptied but pinned sections alone exceed cap → exit 3.
+#   7. --help exits 0 and prints usage; an unknown flag exits nonzero (#701 discipline).
+#   8. More than one marker pair → exit 3, unchanged; curated content between the two
+#      zones is never relocated to the LEDGER (rev0 #868 regression).
+
+set -euo pipefail
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SUT="$SCRIPT_DIR/board-roll.sh"
+
+fail=0
+note() { echo "FAIL: $*" >&2; fail=1; }
+
+WORK="$(mktemp -d)"
+trap 'rm -rf "$WORK"' EXIT
+
+# builds a LIVE board: pinned preamble + roll zone with N dated entries (newest first),
+# each entry padded to be individually large so the cap math is predictable.
+make_board() {  # $1 file  $2 n_entries  $3 with_markers(1/0)  $4 pad_bytes
+    local f=$1 n=$2 markers=$3 pad=$4 i padtxt
+    padtxt=$(head -c "$pad" < /dev/zero | tr '\0' 'x')
+    {
+        echo "# BOARD — LIVE"
+        echo "> pinned protocol blockquote, never rolled."
+        echo
+        echo "## Curated always-current section (pinned)"
+        echo "- this stays no matter what"
+        echo
+        [[ "$markers" == 1 ]] && echo '<!-- BOARD-ROLL:START -->'
+        # newest first (i=n .. 1); oldest (i=1) ends at the bottom
+        for (( i=n; i>=1; i-- )); do
+            echo "### 2026-07-$(printf '%02d' $i) tick number $i"
+            echo "- detail $i $padtxt"
+            echo
+        done
+        [[ "$markers" == 1 ]] && echo '<!-- BOARD-ROLL:END -->'
+    } > "$f"
+    return 0
+}
+
+# ── 1. under cap → no-op ───────────────────────────────────────────────────────
+L="$WORK/live1.md"; G="$WORK/ledger1.md"; : > "$G"
+make_board "$L" 2 1 10
+before=$(cat "$L")
+if ! out=$(bash "$SUT" --live "$L" --ledger "$G" --cap 100000 2>&1); then
+    note "under-cap should exit 0 (got nonzero): $out"
+fi
+[[ "$(cat "$L")" == "$before" ]] || note "under-cap modified LIVE"
+[[ -s "$G" ]] && note "under-cap wrote to LEDGER"
+
+# ── 2. over cap, no markers → exit 3, unchanged ────────────────────────────────
+L="$WORK/live2.md"; G="$WORK/ledger2.md"; : > "$G"
+make_board "$L" 6 0 400
+before=$(cat "$L")
+set +e; bash "$SUT" --live "$L" --ledger "$G" --cap 800 >/dev/null 2>&1; rc=$?; set -e
+[[ "$rc" -eq 3 ]] || note "no-markers over-cap should exit 3 (got $rc)"
+[[ "$(cat "$L")" == "$before" ]] || note "no-markers run modified LIVE (must never guess)"
+
+# ── 3+4. over cap with markers → rolls oldest, LIVE under cap, LEDGER gets them ─
+L="$WORK/live3.md"; G="$WORK/ledger3.md"; echo "# LEDGER" > "$G"
+make_board "$L" 6 1 400          # 6 entries, each ~>400B
+big=$(wc -c < "$L")
+[[ "$big" -ge 2000 ]] || note "fixture too small to test rolling ($big B)"
+if ! out=$(bash "$SUT" --live "$L" --ledger "$G" --cap 2000 2>&1); then
+    note "marker roll should exit 0 when it can get under cap: $out"
+fi
+after=$(wc -c < "$L")
+[[ "$after" -lt 2000 ]] || note "LIVE still >= cap after roll ($after B)"
+# pinned content survives
+grep -q "Curated always-current section" "$L" || note "roll dropped pinned section"
+grep -q 'BOARD-ROLL:START' "$L" || note "roll dropped START marker"
+grep -q 'BOARD-ROLL:END' "$L"   || note "roll dropped END marker"
+# newest entry (07-06) stays; oldest (07-01) is the first to leave
+grep -q "### 2026-07-06 tick number 6" "$L" || note "roll dropped the newest entry"
+grep -q "### 2026-07-01 tick number 1" "$L" && note "oldest entry not rolled out of LIVE"
+# oldest went to LEDGER
+grep -q "### 2026-07-01 tick number 1" "$G" || note "oldest entry not appended to LEDGER"
+grep -q "board-roll:.*rolled from live3.md" "$G" || note "LEDGER missing provenance separator"
+# a rolled entry must not be duplicated (present in exactly one of LIVE/LEDGER)
+if grep -q "### 2026-07-01 tick number 1" "$L"; then note "rolled entry duplicated in LIVE"; fi
+# LEDGER original content preserved
+grep -q "^# LEDGER" "$G" || note "roll clobbered existing LEDGER content"
+
+# ── 5. --dry-run changes nothing ───────────────────────────────────────────────
+L="$WORK/live5.md"; G="$WORK/ledger5.md"; echo "# LEDGER" > "$G"
+make_board "$L" 6 1 400
+before_l=$(cat "$L"); before_g=$(cat "$G")
+out=$(bash "$SUT" --live "$L" --ledger "$G" --cap 2000 --dry-run 2>&1) || note "dry-run exited nonzero: $out"
+echo "$out" | grep -qi "dry run" || note "dry-run did not announce itself"
+echo "$out" | grep -q "would roll" || note "dry-run did not report a plan"
+[[ "$(cat "$L")" == "$before_l" ]] || note "dry-run modified LIVE"
+[[ "$(cat "$G")" == "$before_g" ]] || note "dry-run modified LEDGER"
+
+# ── 6. zone emptied, pinned alone over cap → exit 3 ────────────────────────────
+# cap 120 is below the pinned preamble+footer size (~180B), so even after rolling
+# every zone entry the LIVE file stays over cap → must report the unsatisfiable case.
+L="$WORK/live6.md"; G="$WORK/ledger6.md"; echo "# LEDGER" > "$G"
+make_board "$L" 3 1 50
+set +e; bash "$SUT" --live "$L" --ledger "$G" --cap 120 >/dev/null 2>&1; rc=$?; set -e
+[[ "$rc" -eq 3 ]] || note "unsatisfiable cap should exit 3 (got $rc)"
+
+# ── 7. help exits 0, unknown flag exits nonzero (#701) ─────────────────────────
+if ! out=$(bash "$SUT" --help 2>&1); then note "--help exited nonzero"; fi
+[[ "$out" == Usage:* ]] || note "--help did not print usage"
+bash "$SUT" -h >/dev/null 2>&1 || note "-h exited nonzero"
+if bash "$SUT" --not-a-real-flag >/dev/null 2>&1; then note "unknown flag was accepted"; fi
+if bash "$SUT" --live "$WORK/live3.md" >/dev/null 2>&1; then note "missing --ledger was accepted"; fi
+
+# ── 8. multiple marker pairs → exit 3, unchanged (no cross-zone relocation) ─────
+# Two separately-marked zones with a curated pinned section BETWEEN them. A naive
+# first-START..last-END span would sweep that curated section (and the intermediate
+# markers) into the LEDGER. board-roll must refuse (exit 3) and touch nothing.
+L="$WORK/live8.md"; G="$WORK/ledger8.md"; echo "# LEDGER" > "$G"
+pad8=$(head -c 300 < /dev/zero | tr '\0' 'x')
+{
+    echo "# BOARD — LIVE"
+    echo "> pinned protocol blockquote"
+    echo
+    echo '<!-- BOARD-ROLL:START -->'
+    echo "### 2026-07-10 zone-A newest"
+    echo "- detail A2 $pad8"
+    echo "### 2026-07-09 zone-A oldest"
+    echo "- detail A1 $pad8"
+    echo '<!-- BOARD-ROLL:END -->'
+    echo
+    echo "## Curated-between-zones (pinned — must never move)"
+    echo "- CANARY-BETWEEN keep me"
+    echo
+    echo '<!-- BOARD-ROLL:START -->'
+    echo "### 2026-07-08 zone-B newest"
+    echo "- detail B2 $pad8"
+    echo "### 2026-07-07 zone-B oldest"
+    echo "- detail B1 $pad8"
+    echo '<!-- BOARD-ROLL:END -->'
+} > "$L"
+before8=$(cat "$L")
+set +e; bash "$SUT" --live "$L" --ledger "$G" --cap 80 >/dev/null 2>&1; rc=$?; set -e
+[[ "$rc" -eq 3 ]] || note "multi-pair board should exit 3 (got $rc)"
+[[ "$(cat "$L")" == "$before8" ]] || note "multi-pair run modified LIVE (must never guess across zones)"
+grep -q "CANARY-BETWEEN keep me" "$L" || note "multi-pair run relocated curated between-zones content"
+grep -q "CANARY-BETWEEN" "$G" && note "curated between-zones content leaked into LEDGER"
+
+if [[ "$fail" -eq 0 ]]; then
+    echo "board-roll regression passed (8 groups)"
+fi
+exit "$fail"

packages/mosaic/src/mutator-gate/mutator-gate.acceptance.spec.ts

@@ -661,13 +661,27 @@ describe('whole mutator-class lease gate', () => {
   test('observer revocation and monotonic TTL expiry deny the next mutator', async () => {
     const { socket } = await startBroker();
     const sessionId = await register(socket);
-    const pending = await beginVerification(socket, sessionId, 'claude', 1, 1);
-    await promote(socket, sessionId, pending.receipt_challenge!);
 
+    // Establish the lease with a normal (non-racing) TTL first and prove it
+    // authorizes. This "still valid" check is setup, not a TTL-expiry
+    // assertion, so it must not share a lease with a 1-second TTL: on a
+    // contended push-CI host, scheduling delay alone between promote() and
+    // this authorize() call can consume that entire 1-second margin and
+    // spuriously deny it (CI#1945). Using a generous TTL here removes that
+    // real-time race without touching lease-gate security semantics.
+    const pending = await beginVerification(socket, sessionId, 'claude');
+    await promote(socket, sessionId, pending.receipt_challenge!);
     expect(await authorize(socket, sessionId, 'claude', 'Bash')).toMatchObject({
       ok: true,
       decision: 'allow',
     });
+
+    // A dedicated, isolated short-TTL lease drives the deliberate monotonic
+    // expiry demonstration below. It is never used for anything but the
+    // wait-then-expire assertion, so there is no setup work racing its
+    // 1-second window.
+    const shortLived = await beginVerification(socket, sessionId, 'claude', 1, 1, 2);
+    await promote(socket, sessionId, shortLived.receipt_challenge!);
     await new Promise((resolve) => setTimeout(resolve, 1_100));
     expect(await authorize(socket, sessionId, 'claude', 'Bash')).toMatchObject({
       ok: false,
@@ -675,7 +689,7 @@ describe('whole mutator-class lease gate', () => {
       decision: 'deny',
     });
 
-    const refreshed = await beginVerification(socket, sessionId, 'claude', 1, 300, 2);
+    const refreshed = await beginVerification(socket, sessionId, 'claude', 1, 300, 3);
     await promote(socket, sessionId, refreshed.receipt_challenge!);
     expect(
       await request(socket, {

skills/glpi-create/SKILL.md

@@ -0,0 +1,50 @@
+# Skill: glpi-create — Open a New GLPI Ticket
+
+> Create a new GLPI helpdesk ticket. Mutates GLPI — confirm the details before running.
+
+## When to use
+
+- Logging a new incident or request that should live in the helpdesk queue.
+
+## Required information
+
+- **title** — short subject line.
+- **content** — description of the issue / request.
+
+## Optional
+
+- **priority** — `1`=VeryLow, `2`=Low, `3`=Medium (default), `4`=High, `5`=VeryHigh, `6`=Major.
+- **type** — `1`=Incident (default), `2`=Request.
+
+## Command
+
+Wraps the existing tooling:
+
+```bash
+~/.config/mosaic/tools/glpi/ticket-create.sh \
+  -t "<title>" \
+  -c "<content>" \
+  [-p <priority>] \
+  [-y <type>] \
+  [-f json]
+```
+
+Example:
+
+```bash
+~/.config/mosaic/tools/glpi/ticket-create.sh \
+  -t "Paint-area camera install" \
+  -c "Ordered 2 cameras for Paint and stock; schedule mounting + NVR config." \
+  -p 3 -y 2
+```
+
+## After creating
+
+- Note the returned **ticket ID** — you'll need it for **[[glpi-followup]]** and
+  **[[glpi-solve]]**.
+- If it should also be tracked as brain work, add a matching task (see the `add-task` skill).
+
+## Guardrails
+
+- Confirm title/content/priority with the user before creating — a ticket is outward-facing.
+- Never echo GLPI tokens.

skills/glpi-followup/SKILL.md

@@ -0,0 +1,56 @@
+# Skill: glpi-followup — Add a Followup to a GLPI Ticket
+
+> Post a followup (comment / progress note / resolution writeup) to a GLPI ticket.
+> This documents work but does **not** change the ticket status — to close a ticket
+> out, follow with **[[glpi-solve]]** to set status to Solved.
+
+## When to use
+
+- Recording progress, a decision, or a root-cause/resolution note on a ticket.
+- The documentation step that usually precedes closing a ticket out (`glpi-solve`).
+
+## Critical quirk
+
+Use the **top-level `/ITILFollowup` endpoint**, NOT `/Ticket/<id>/ITILFollowup`. The
+sub-resource path returns permission errors even with a Super-Admin profile.
+
+## Procedure
+
+### 1. Session + creds
+
+```bash
+SESSION=$(~/.config/mosaic/tools/glpi/session-init.sh -q)
+source ~/.config/mosaic/tools/_lib/credentials.sh && load_credentials glpi
+```
+
+### 2. Post the followup
+
+```bash
+TICKET_ID=<id>
+CONTENT="<the followup text>"
+curl -sk -X POST "${GLPI_URL}/ITILFollowup" \
+  -H "App-Token: $GLPI_APP_TOKEN" \
+  -H "Session-Token: $SESSION" \
+  -H "Content-Type: application/json" \
+  -d "$(jq -n --argjson id "$TICKET_ID" --arg c "$CONTENT" \
+        '{input:{itemtype:"Ticket", items_id:$id, content:$c}}')"
+```
+
+Expect HTTP 201. Building the payload with `jq` keeps quotes/newlines in the content safe.
+
+### 3. Long or multi-paragraph content
+
+Write the note to a file first, then read it into the payload:
+
+```bash
+curl -sk -X POST "${GLPI_URL}/ITILFollowup" \
+  -H "App-Token: $GLPI_APP_TOKEN" -H "Session-Token: $SESSION" \
+  -H "Content-Type: application/json" \
+  -d "$(jq -n --argjson id "$TICKET_ID" --rawfile c /path/to/note.md \
+        '{input:{itemtype:"Ticket", items_id:$id, content:$c}}')"
+```
+
+## Guardrails
+
+- Never echo the GLPI app/user/session tokens.
+- A followup alone leaves the ticket open. If the work is done, run **[[glpi-solve]]** next.

skills/glpi-list/SKILL.md

@@ -0,0 +1,57 @@
+# Skill: glpi-list — Query GLPI Tickets
+
+> Quick lookups of GLPI helpdesk tickets by status or recency. Read-only.
+
+## When to use
+
+- "What tickets are open / pending?" · "Show recent tickets" · finding a ticket ID
+  before running **[[glpi-followup]]** or **[[glpi-solve]]**.
+
+## Command
+
+Wraps the existing tooling:
+
+```bash
+GLPI=~/.config/mosaic/tools/glpi
+
+# Most recent tickets (default 50, newest first)
+"$GLPI/ticket-list.sh"
+
+# Filter by status: new | processing | pending | solved | closed
+"$GLPI/ticket-list.sh" -s pending
+
+# JSON output (for parsing / piping to jq) and a custom limit
+"$GLPI/ticket-list.sh" -s processing -f json -l 20
+```
+
+Status IDs: 1 New · 2/3 Processing · 4 Pending · 5 Solved · 6 Closed.
+
+## Details lookup for one ticket
+
+When you have an ID and want the full record:
+
+```bash
+SESSION=$(~/.config/mosaic/tools/glpi/session-init.sh -q)
+source ~/.config/mosaic/tools/_lib/credentials.sh && load_credentials glpi
+curl -sk "${GLPI_URL}/Ticket/<id>?expand_dropdowns=true" \
+  -H "App-Token: $GLPI_APP_TOKEN" -H "Session-Token: $SESSION" \
+  | jq '{id, name, status, date, date_mod}'
+
+# Followups on a ticket
+curl -sk "${GLPI_URL}/Ticket/<id>/ITILFollowup" \
+  -H "App-Token: $GLPI_APP_TOKEN" -H "Session-Token: $SESSION" \
+  | jq '.[] | {date, content}'
+```
+
+(Reading followups via the sub-resource is fine — only _creating_ them requires the
+top-level `/ITILFollowup` endpoint. See **[[glpi-followup]]**.)
+
+## Present to user
+
+Group by status, one line per ticket: `#<id> · <title> · <status> · <last-modified>`.
+Use neutral phrasing — no "OVERDUE"/"URGENT".
+
+## Guardrails
+
+- Read-only. Never echo GLPI tokens.
+- To sync tickets into brain data instead, use `python tools/sync_glpi.py` (not this skill).

skills/glpi-solve/SKILL.md

@@ -0,0 +1,96 @@
+# Skill: glpi-solve — Close Out a GLPI Ticket
+
+> Properly close out a completed GLPI helpdesk ticket. Completing the work is not
+> enough — the ticket **status must be set to "Solved"**, which is what triggers
+> GLPI's config-driven auto-close. Posting a resolution followup documents the work
+> but does **not** change status, so a ticket left at Solved-less status stays open.
+
+## When to use
+
+- Any time work on a GLPI ticket is finished and it should be closed out.
+- After posting a root-cause / resolution writeup as an `/ITILFollowup`.
+- During a cleanup sweep of tickets that are done in reality but still open in GLPI.
+
+## The rule (from an operator, 2026-07-20)
+
+**"Solved" is the correct terminal state to set — not "Closed."** GLPI is configured
+to auto-close Solved tickets after its delay. If you only post a followup and never set
+status, the ticket sits open (this bit us on a real incident where resolution followups
+were posted but status was never advanced, leaving tickets open, which the operator had
+to mark Solved by hand).
+
+Close-out = **followup (optional but preferred) + set status to Solved.**
+
+## GLPI status IDs
+
+| ID    | Status                |                                              |
+| ----- | --------------------- | -------------------------------------------- |
+| 1     | New                   |                                              |
+| 2     | Processing (assigned) |                                              |
+| 3     | Processing (planned)  |                                              |
+| 4     | Pending / Waiting     |                                              |
+| **5** | **Solved**            | ← set this on close-out                      |
+| 6     | Closed                | ← happens automatically; do not set manually |
+
+## Procedure
+
+### 1. Get a session token
+
+```bash
+SESSION=$(~/.config/mosaic/tools/glpi/session-init.sh -q)
+source ~/.config/mosaic/tools/_lib/credentials.sh && load_credentials glpi
+```
+
+### 2. (Preferred) Post the resolution followup
+
+Use the **top-level `/ITILFollowup` endpoint** — the `/Ticket/<id>/ITILFollowup`
+sub-resource returns permission errors even as Super-Admin (known GLPI quirk).
+
+```bash
+TICKET_ID=<id>
+curl -sk -X POST "${GLPI_URL}/ITILFollowup" \
+  -H "App-Token: $GLPI_APP_TOKEN" \
+  -H "Session-Token: $SESSION" \
+  -H "Content-Type: application/json" \
+  -d "{\"input\":{\"itemtype\":\"Ticket\",\"items_id\":${TICKET_ID},\"content\":\"<resolution summary>\"}}"
+```
+
+### 3. Set status to Solved (the step that actually closes it out)
+
+```bash
+curl -sk -X PUT "${GLPI_URL}/Ticket/${TICKET_ID}" \
+  -H "App-Token: $GLPI_APP_TOKEN" \
+  -H "Session-Token: $SESSION" \
+  -H "Content-Type: application/json" \
+  -d "{\"input\":{\"id\":${TICKET_ID},\"status\":5}}"
+```
+
+Expect HTTP 200/201. GLPI will auto-close it later per its config — leave status at 5.
+
+### 4. Verify
+
+```bash
+curl -sk "${GLPI_URL}/Ticket/${TICKET_ID}?expand_dropdowns=true" \
+  -H "App-Token: $GLPI_APP_TOKEN" -H "Session-Token: $SESSION" \
+  | jq '{id, name, status}'
+```
+
+`status` should read `Solved` (or `5`).
+
+## Optional: sweep for done-but-open tickets
+
+List tickets still open (New/Processing/Pending) to spot ones whose work is actually
+finished but were never marked Solved:
+
+```bash
+~/.config/mosaic/tools/glpi/ticket-list.sh -s processing -f table
+~/.config/mosaic/tools/glpi/ticket-list.sh -s pending    -f table
+```
+
+Review each; for any that are genuinely resolved, run steps 2–3.
+
+## Guardrails
+
+- Read-only until you intend to close — confirm the ticket is actually done first.
+- Never echo the GLPI app/user/session tokens.
+- Set **Solved (5)**, never Closed (6) — auto-close owns that transition.

skills/glpi-sweep/SKILL.md

@@ -0,0 +1,62 @@
+# Skill: glpi-sweep — Find Done-But-Open Tickets
+
+> Read-only sweep for tickets that are finished in reality but still sitting open in
+> GLPI (never moved to Solved). Surfaces the exact miss an operator caught on 2026-07-20
+> (a real incident where an affected ticket had resolution followups posted but was left
+> open). For each one that's genuinely done, close it out with **[[glpi-solve]]**.
+
+## When to use
+
+- Periodic hygiene pass (e.g. before a weekly update or month-end).
+- After a burst of ticket work, to catch any you resolved-in-followup but never Solved.
+
+## Why this exists
+
+Posting an `/ITILFollowup` documents work but does **not** change status. Tickets only
+auto-close once set to **Solved (status 5)**. Anything left at New/Processing/Pending
+stays open indefinitely. This sweep finds those.
+
+## Procedure
+
+### 1. List still-open tickets by status
+
+```bash
+GLPI=~/.config/mosaic/tools/glpi
+"$GLPI/ticket-list.sh" -s new        -f table
+"$GLPI/ticket-list.sh" -s processing -f table
+"$GLPI/ticket-list.sh" -s pending    -f table
+```
+
+(GLPI status IDs: 1 New · 2/3 Processing · 4 Pending · 5 Solved · 6 Closed.)
+
+### 2. Triage
+
+For each open ticket, judge whether the underlying work is actually finished — check
+its latest followups and cross-reference brain tasks / recent work. Read-only here;
+change nothing yet.
+
+Reasonable "probably done" signals:
+
+- A resolution/root-cause followup already posted, but status never advanced.
+- The related brain task is `done`, or the fix shipped and was confirmed.
+- Requester confirmed resolution but the ticket was never Solved.
+
+### 3. Present the candidates
+
+List them for review before touching anything — never bulk-solve blindly:
+
+```
+Open tickets that look resolved:
+- #<id> "<title>" — <why it looks done> → glpi-solve?
+```
+
+### 4. Close out the confirmed ones
+
+For each ticket the user (or clear evidence) confirms is done, run **[[glpi-solve]]**
+(optionally **[[glpi-followup]]** first if a closing note is warranted).
+
+## Guardrails
+
+- Read-only until a ticket is confirmed done — do not auto-solve on a guess.
+- Never echo GLPI tokens.
+- Set **Solved (5)**, never Closed (6) — GLPI auto-close owns that transition.