MCP.Market

gstack-upgrade skill

by garrytan·garrytan/gstack·142k stars·MIT

Upgrade gstack to the latest version.

C70/100content scan

Is the gstack-upgrade skill safe?

Read the findings before you install it. We read 17 files in the folder on 2026-11-01.

  • highSKILL.md:180

    Downloads a script and runs it in one step, so what runs is whatever that server sends that day. Common for installers, and still worth a look at the address.

    On `BUN_VERSION_UNREADABLE`, STOP: `bun --version` did not print a version gstack can read, so `bun upgrade` may not run either; tell the user to reinstall Bun from https://bun.sh (`curl -fsSL http...

Install the gstack-upgrade skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description.

git clone --depth 1 https://github.com/garrytan/gstack.git /tmp/gstack
mkdir -p ~/.claude/skills
cp -r /tmp/gstack/gstack-upgrade ~/.claude/skills/gstack-upgrade
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

<!-- AUTO-GENERATED from SKILL.md.tmpl — do not edit directly --> <!-- Regenerate: bun run gen:skill-docs -->

/gstack-upgrade

Upgrade gstack to the latest version and show what's new.

Inline upgrade flow

This section is referenced by all skill preambles when they detect UPGRADE_AVAILABLE.

Step 1: Ask the user (or auto-upgrade)

First, check if auto-upgrade is enabled:

_AUTO=""
[ "${GSTACK_AUTO_UPGRADE:-}" = "1" ] && _AUTO="true"
[ -z "$_AUTO" ] && _AUTO=$(~/.claude/skills/gstack/bin/gstack-config get auto_upgrade 2>/dev/null || true)
echo "AUTO_UPGRADE=$_AUTO"

If AUTOUPGRADE=true or AUTOUPGRADE=1: Skip AskUserQuestion. Log "Auto-upgrading gstack v{old} → v{new}..." and proceed directly to Step 2. On setup failure, follow Step 4's install-specific recovery: vendored installs restore their backup; git installs stop with the pre-upgrade commit recorded, without a destructive reset. Never claim restoration unless it actually succeeded.

Otherwise, use AskUserQuestion:

Four fields (UPGRADEAVAILABLE + , main moved): ask "gstack main is {n} commit(s)** past v{old} (at {sha}). Upgrade now?"

  • Question: "gstack v{new} is available (you're on v{old}). Upgrade now?"
  • Options: ["Yes, upgrade now", "Always keep me up to date", "Not now", "Never ask again"]

If "Yes, upgrade now": Proceed to Step 2.

If "Always keep me up to date":

~/.claude/skills/gstack/bin/gstack-config set auto_upgrade true

Tell user: "Auto-upgrade enabled. Future updates will install automatically." Then proceed to Step 2.

If "Not now": Write snooze state with escalating backoff (first snooze = 24h, second = 48h, third+ = 1 week), then continue with the current skill. Do not mention the upgrade again.

GSTACK_STATE_ROOT=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT); : "${GSTACK_STATE_ROOT:?gstack-paths failed; reinstall with ./setup or /gstack-upgrade}"
_SNOOZE_FILE="$GSTACK_STATE_ROOT/update-snoozed"
_REMOTE_VER="{new}"
_CUR_LEVEL=0
if [ -f "$_SNOOZE_FILE" ]; then
  _SNOOZED_VER=$(awk '{print $(1)}' "$_SNOOZE_FILE")
  if [ "$_SNOOZED_VER" = "$_REMOTE_VER" ]; then
    _CUR_LEVEL=$(awk '{print $(2)}' "$_SNOOZE_FILE")
    case "$_CUR_LEVEL" in *[!0-9]*) _CUR_LEVEL=0 ;; esac
  fi
fi
_NEW_LEVEL=$((_CUR_LEVEL + 1))
[ "$_NEW_LEVEL" -gt 3 ] && _NEW_LEVEL=3
echo "$_REMOTE_VER $_NEW_LEVEL $(date +%s)" > "$_SNOOZE_FILE"

Note: {new} is the remote version from the UPGRADE_AVAILABLE output — substitute it from the update check result. On a four-field line use the SHA (fourth field) as {new}, so the snooze ends when main moves.

Tell user the snooze duration: "Next reminder in 24h" (or 48h or 1 week, depending on level). Tip: "Set auto_upgrade: true in ~/.gstack/config.yaml for automatic upgrades."

If "Never ask again":

~/.claude/skills/gstack/bin/gstack-config set update_check false

Tell user: "Update checks disabled. Run ~/.claude/skills/gstack/bin/gstack-config set update_check true to re-enable." Continue with the current skill.

Step 2: Detect install type

if [ -d "$HOME/.claude/skills/gstack/.git" ]; then
  INSTALL_TYPE="global-git"
  INSTALL_DIR="$HOME/.claude/skills/gstack"
elif [ -d "$HOME/.gstack/repos/gstack/.git" ]; then
  INSTALL_TYPE="global-git"
  INSTALL_DIR="$HOME/.gstack/repos/gstack"
elif [ -d ".claude/skills/gstack/.git" ]; then
  INSTALL_TYPE="local-git"
  INSTALL_DIR=".claude/skills/gstack"
elif [ -d ".agents/skills/gstack/.git" ]; then
  INSTALL_TYPE="local-git"
  INSTALL_DIR=".agents/skills/gstack"
elif [ -d ".claude/skills/gstack" ]; then
  INSTALL_TYPE="vendored"
  INSTALL_DIR=".claude/skills/gstack"
elif [ -d "$HOME/.claude/skills/gstack" ]; then
  INSTALL_TYPE="vendored-global"
  INSTALL_DIR="$HOME/.claude/skills/gstack"
elif _SRC=$(~/.claude/skills/gstack/bin/gstack-paths --get GSTACK_STATE_ROOT 2>/dev/null) && _SRC=$(awk -F '\t' '$(1) == "claude" && $(6) != "-" { print $(6); exit }' "$_SRC/installs.tsv" 2>/dev/null) && [ -d "$_SRC/.git" ]; then
  # The install registry names the checkout setup activated for this host (e.g. a ~/gstack clone).
  INSTALL_TYPE="global-git"
  INSTALL_DIR="$_SRC"
else
  echo "ERROR: gstack not found"
  exit 1
fi
INSTALL_DIR=$(cd -- "$INSTALL_DIR" && pwd -P) || { echo "ERROR: cannot

The install type and the absolute directory printed above are used in all subsequent steps. Carry INSTALLTYPE, INSTALLDIR, OLDVERSION, and later UPSTREAMREF and NEWVERSION forward explicitly: tool calls may run in fresh shells, so start each later block by assigning them from the captured output (for example INSTALLDIR=/abs/path/printed/above). Do not rely on a prior call's working directory or shell variables. Every block that changes files first checks INSTALL_DIR and stops before any git command when it is unset, empty, missing, unreadable, or not gstack's own checkout.

Step 3: Save old version

Use the install directory from Step 2's output below:

OLD_VERSION=$(cat "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}/VERSION" 2>/dev/null || echo "unknown")
echo "OLD_VERSION=$OLD_VERSION"

Step 4: Upgrade

Use the install type and directory detected in Step 2:

For git installs (global-git, local-git):

Fast-forward first, through bin/gstack-ff-update, the same code the session-update auto-upgrade uses: local edits that merge cleanly ride over the update, and an edit that conflicts (or a staged change, or an unmerged file) is saved in a rescue ref instead of being left as conflict markers:

cd -- "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}" || exit 1
{ [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ] && _p=$(git rev-parse --show-prefix 2>/dev/null) && [ -z "$_p" ] && [ "$(git rev-parse --is-inside-work-tree 2>/dev/null)" = true ]; } || { echo "ERROR: $INSTALL_DIR is not a gstack checkout; nothing was changed. Re-run Step 2." >&2; exit 1; }
git fetch origin
# Channel ref (docs/releases.md): origin/stable for a team install, else origin/main
_CH=$(bash bin/gstack-update-channel.sh --fetch-ref 2>&1) || _CH=origin/main
echo "$_CH"; UPSTREAM_REF=$(printf '%s\n' "$_CH" | tail -1)
echo "UPSTREAM_REF=${UPSTREAM_REF:=origin/main}"
# Same pre-advance check as the auto-updater: the incoming release's Bun floor,
# judged by bash whatever this tool shell is (zsh does not word-split).
HOLD=$(UPSTREAM_REF="$UPSTREAM_REF" bash -c '. bin/gstack-bun-version.sh && gstack_bun_incoming_hold . "$UPSTREAM_REF"' 2>/dev/null) && { case "$HOLD" in bun-version-unreadable*) echo "BUN_VERSION_UNREADABLE: $HOLD; nothing was changed" >&2 ;; *) echo "BUN_TOO_OLD: $HOLD; nothing was changed" >&2 ;; esac; exit 1; }
PRE_UPGRADE_COMMIT=$(git rev-parse

Setup's --refresh-registered refreshes every install this checkout registered (each host, global and this project's), prints the source first, and ends with an Upgrade summary of one row per install. Relay those rows as printed. A failed row means that host kept its previous install; give its retry command and never say every host was refreshed. A skipped row belongs to another checkout or project; pass on its command, do not run it.

If the output ends with FFOK, the upgrade is done — skip the fallback below entirely. When the line before it reads rescued refs/gstack/rescue/ unmerged=<0|1>, tell the user their conflicting local edits were saved there and relay the Inspect, Export and Delete commands the script printed (export into a scratch checkout with git apply -3, never stash apply in the install; unmerged=1 means the rescue needs manual resolution). On FFSTOPPED, STOP and relay the reason: nothing was changed. On AHEADOFSTABLE, STOP: the checkout is already past the latest stable release, so nothing changes until the next cut (gstack-config set updatechannel main follows every merge). On a fetching origin/stable failed line above, STOP and relay it; nothing was changed. ff-busy means another gstack update is running (retry in a minute); autostash-unrecoverable names what is in the way (an untracked file the release adds, or another git command); index-unreadable means the checkout's index is damaged. On BUNTOOOLD, STOP: tell the user to run bun upgrade, then /gstack-upgrade again. On BUNVERSIONUNREADABLE, STOP: bun --version did not print a version gstack can read, so bun upgrade may not run either; tell the user to reinstall Bun from https://bun.sh (curl -fsSL https://bun.sh/install | bash in Git Bash or a macOS/Linux terminal), check that bun --version prints a version such as 1.4.2, then run /gstack-upgrade again (docs/troubleshooting.md#bun-version-unreadable). On SETUPFAILED, STOP; keep user changes and report the recovery commit. There is no .bak on the git path. Do not enter the divergence fallback merely because setup failed. Enter it only on FF_REFUSED (local commits or a diverged history); network/auth failures stop for repair, not reset.

Fallback (ff-only refused — local commits or divergence). git reset --hard DESTROYS things: a clean tree with unpushed local commits still loses those commits. Gate it:

in $INSTALL_DIR.

  1. Run git status --porcelain and git rev-list "${UPSTREAM_REF:?}"..HEAD --oneline

below without asking.

  1. If BOTH are empty, the reset is provably safe — run the fallback block

exactly what will be discarded: each dirty file and each unpushed commit by hash + subject. Options: A) Discard them and upgrade (reset) — requires the explicit letter; B) Abort the upgrade so the user can rescue their work first (recommended when local commits exist). Never proceed on a vague reply.

  1. Otherwise ask via AskUserQuestion (one-way door — destructive), listing
cd -- "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}" || exit 1
{ [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ] && _p=$(git rev-parse --show-prefix 2>/dev/null) && [ -z "$_p" ] && [ "$(git rev-parse --is-inside-work-tree 2>/dev/null)" = true ]; } || { echo "ERROR: $INSTALL_DIR is not a gstack checkout; nothing was changed. Re-run Step 2." >&2; exit 1; }
# Regenerable render-footprint dirt (setup re-renders it) never goes into the stash.
git checkout -- 'SKILL.md' '*/SKILL.md' '*/sections/*.md' 2>/dev/null || true
STASH_OUTPUT=$(git stash 2>&1)
git reset --hard "${UPSTREAM_REF:?UPSTREAM_REF is not set: substitute the value Step 4 printed}"
./setup --refresh-registered

If $STASH_OUTPUT contains "Saved working directory", warn the user: "Note: local changes were stashed (any modified generated SKILL.md/sections files were discarded first — they regenerate on setup). Run git stash pop in the skill directory to restore your own changes."

For vendored installs (vendored, vendored-global):

(cd -- "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}" && [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ]) || { echo "ERROR: INSTALL_DIR=${INSTALL_DIR:-} is not a gstack install; nothing was changed. Re-run Step 2." >&2; exit 1; }
# A stale .bak from a previously crashed upgrade would make the mv below NEST
# the live install inside it and the failure-restore arm would "restore" the
# stale backup. It may also be the only good copy from that crashed run —
# abort and let the human inspect, never delete it silently.
[ -e "$INSTALL_DIR.bak" ] && { echo "ERROR: stale backup exists at $INSTALL_DIR.bak (from a previous failed upgrade?) — inspect it, salvage/remove it, then re-run." >&2; exit 1; }
TMP_DIR=$(mktemp -d "${TMPDIR:-/tmp}/gstack-upgrade.XXXXXX") || { echo "ERROR: mktemp failed — aborting upgrade (install untouched)." >&2; exit 1; }
git clone --depth 1 https://github.com/garrytan/gstack.git "$TMP_DIR/gstack" || { echo "ERROR: clone failed — aborting upgrade (install untouched)." >&2; rm -rf "$TMP_DIR"; exit 1; }
mv "$INSTALL_DIR" "$INSTALL_DIR.bak" || { rm -rf "$TMP_DIR"; exit 1; }
if mv "$TMP_DIR/gstack" "$INSTALL_DIR"; then

Step 4.5: Handle local vendored copy

Use the install directory from Step 2. Check if there's also a local vendored copy, and whether team mode is active:

_RESOLVED_PRIMARY=$(cd -- "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}" && pwd -P) || exit 1
_ROOT=$(git rev-parse --show-toplevel 2>/dev/null)
LOCAL_GSTACK=""
if [ -n "$_ROOT" ] && [ -d "$_ROOT/.claude/skills/gstack" ]; then
  _RESOLVED_LOCAL=$(cd "$_ROOT/.claude/skills/gstack" && pwd -P)
  if [ "$_RESOLVED_LOCAL" != "$_RESOLVED_PRIMARY" ]; then
    LOCAL_GSTACK="$_ROOT/.claude/skills/gstack"
  fi
fi
_TEAM_MODE=$(~/.claude/skills/gstack/bin/gstack-config get team_mode 2>/dev/null || echo "false")
echo "LOCAL_GSTACK=$LOCAL_GSTACK"
echo "TEAM_MODE=$_TEAM_MODE"

If LOCALGSTACK is non-empty AND TEAMMODE is true: Remove the vendored copy. Team mode uses the global install as the single source of truth.

(cd -- "${LOCAL_GSTACK:?LOCAL_GSTACK is not set: re-run the detection block above and substitute the printed path}" && [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ]) || { echo "ERROR: LOCAL_GSTACK=${LOCAL_GSTACK:-} is not a gstack copy; nothing was changed." >&2; exit 1; }
_ROOT=${LOCAL_GSTACK%/.claude/skills/gstack}
{ [ "$_ROOT" != "$LOCAL_GSTACK" ] && cd -- "${_ROOT:?}" && _p=$(git rev-parse --show-prefix 2>/dev/null) && [ -z "$_p" ] && [ "$(git rev-parse --is-inside-work-tree 2>/dev/null)" = true ]; } || { echo "ERROR: $LOCAL_GSTACK is not the vendored copy at the root of a git repository; nothing was changed." >&2; exit 1; }
git rm -r --cached .claude/skills/gstack/ 2>/dev/null || true
if ! grep -qF '.claude/skills/gstack/' .gitignore 2>/dev/null; then
  echo '.claude/skills/gstack/' >> .gitignore
fi
rm -rf "$LOCAL_GSTACK"

Tell user: "Removed vendored copy at $LOCAL_GSTACK (team mode active — global install is the source of truth). Commit the .gitignore change when ready."

If LOCALGSTACK is non-empty AND TEAMMODE is NOT true: Update it by copying from the freshly-upgraded primary install (same approach as README vendored install):

(cd -- "${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}" && [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ]) || { echo "ERROR: INSTALL_DIR=${INSTALL_DIR:-} is not a gstack install; nothing was changed." >&2; exit 1; }
(cd -- "${LOCAL_GSTACK:?LOCAL_GSTACK is not set: re-run the detection block above and substitute the printed path}" && [ -f VERSION ] && [ -f setup ] && [ -f bin/gstack-config ]) || { echo "ERROR: LOCAL_GSTACK=${LOCAL_GSTACK:-} is not a gstack copy; nothing was changed." >&2; exit 1; }
[ -e "$LOCAL_GSTACK.bak" ] && { echo "ERROR: stale vendored backup; inspect it before retrying." >&2; exit 1; }
mv "$LOCAL_GSTACK" "$LOCAL_GSTACK.bak" || exit 1
if cp -Rf "$INSTALL_DIR" "$LOCAL_GSTACK" && rm -rf "$LOCAL_GSTACK/.git" && (cd "$LOCAL_GSTACK" && ./setup --refresh-registered); then
  rm -rf "$LOCAL_GSTACK.bak"
  echo "LOCAL_SYNC_OK"
else
  rm -rf "$LOCAL_GSTACK"
  mv "$LOCAL_GSTACK.bak" "$LOCAL_GSTACK" || { echo "ERROR: restore failed; backup retained." >&2; exit 1; }
  echo "ERROR: sync failed; previous vendored copy restored." >&2
  exit 1
fi

Only on LOCALSYNCOK, tell user: "Also updated vendored copy at $LOCAL_GSTACK — commit .claude/skills/gstack/ when you're ready." Otherwise stop and report the recovery outcome; do not continue migrations or announce success.

Step 4.75: Run version migrations

After ./setup, run the migrations in gstack-upgrade/migrations/ (idempotent v{VERSION}.sh scripts; CONTRIBUTING.md says how to add one) newer than the old version: they fix state ./setup alone cannot (stale config, orphaned files).

MIGRATIONS_DIR="${INSTALL_DIR:?INSTALL_DIR is not set: re-run Step 2 and substitute the printed path}/gstack-upgrade/migrations"
: "${OLD_VERSION:?OLD_VERSION is not set: substitute the value Step 3 printed}"
if [ -d "$MIGRATIONS_DIR" ]; then
  for migration in $(find "$MIGRATIONS_DIR" -maxdepth 1 -name 'v*.sh' -type f 2>/dev/null | sort -V); do
    m_ver="$(basename "$migration" .sh | sed 's/^v//')"
    if [ "$OLD_VERSION" != "unknown" ] && [ "$(printf '%s\n%s' "$OLD_VERSION" "$m_ver" | sort -V | head -1)" = "$OLD_VERSION" ] && [ "$OLD_VERSION" != "$m_ver" ]; then
      echo "Running migration $m_ver..."
      # Without GSTACK_INSTALL_DIR a migration that cleans the install assumes ~/.claude/skills/gstack.
      GSTACK_INSTALL_DIR="$INSTALL_DIR" bash "$migration" || echo "  Warning: migration $m_ver had errors (non-fatal)"
    fi
  done
fi

More skills from garrytan/gstack

  • AalphaFixture dispatcher with a mode table and forced-read references.
  • AautoplanAuto-review pipeline — runs the CEO, design, eng (architecture, edge cases, failure modes, tests) and DX reviews (lean run: one reviewer), auto-deciding. (gstack)
  • AbenchmarkWeb performance regression detection. (gstack)
  • Abenchmark-modelsCross-model benchmark for gstack skills. (gstack)
  • AbetaClean fixture tool skill with no forced reads and no mode table.
  • FbrowseDrive a real browser through Aside: open a page, read it, click through a flow, take screenshots, check console errors. (gstack)
  • AcanaryPost-deploy canary monitoring. (gstack)
  • AcaptainRun a crew of coding agents from one controller (also /crew): lanes send READY blocks and never merge; the controller merges through trains, keeps the digest, relays approvals, fixes flakes daily. (gstack)
  • FcarefulSafety guardrails for destructive commands. (gstack)
  • AcodexOpenAI Codex CLI wrapper — three modes. (gstack)
  • Acontext-restoreRestore working context saved earlier by /context-save. (gstack)
  • Acontext-saveSave working context. (gstack)

All agent skills → · MCP servers