Commit Graph
8 Commits
Author SHA1 Message Date
Garry Tanandcapy-ai[bot] 6b327148e3 Stop gstack-uninstall deleting unpushed or kept work under any folder it removes (GST-2303) (#4282)
* Stop gstack-uninstall deleting unpushed or kept work under any folder it removes (GST-2303)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Keep repos behind a symlinked .gstack-worktrees, remove stable-channel installs, and read worktrees on git before 2.36 in gstack-uninstall (GST-2303)

Review-Finding: p1/red-team#1 reader=red-team
Review-Finding: p1/red-team#2 reader=red-team
Review-Finding: p1/red-team#3 reader=red-team
Review-Finding: p1/specialist-batched#1 reader=testing
Review-Finding: p1/specialist-batched#2 reader=testing
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Count a FETCH_HEAD commit as pushed only when it came from one of the checkout's remotes in gstack-uninstall (GST-2303)

Review-Finding: p2/specialist-batched#1 reader=security
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Read remote URLs with git remote get-url so url.insteadOf rewrites match FETCH_HEAD in gstack-uninstall (GST-2303)

Review-Finding: fc/fix-check#1 reader=fix-check
Review-Finding: fc/fix-check#2 reader=fix-check
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Refuse to uninstall on git older than 2.36 instead of reading worktrees without -z (GST-2303)

Review-Finding: v11/outside-voice#1 reader=outside-voice
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* List the git 2.36 uninstall refusal in troubleshooting (GST-2303)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-11-10 09:34:35 +00:00
Garry Tanandcapy-ai[bot] 6c23638d6d Stop gstack-uninstall deleting nested worktrees, locked worktrees and a .gstack repo (GST-2268) (#4267)
* Stop gstack-uninstall deleting nested worktrees, locked worktrees and a .gstack repo (GST-2268)

The worktree check now walks every checkout under .gstack-worktrees/,
nested ones included, and keeps a submodule of a clean lane covered by
that lane's status. A locked worktree stops the uninstall even with
--force-worktrees. A .gstack/ that holds its own git repository is kept
with a message. The check runs again right before .gstack-worktrees/ is
deleted, so work saved while the prompt waited is kept (exit 2). The
manual steps in docs/uninstall.md no longer rm -rf the per-project
folders.

test/uninstall.test.ts is already excluded from the Windows subset by
scripts/lib/windows-curation.ts (it spawns a bin/ shebang script).

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Keep a lane's submodule when it holds commits no remote-tracking branch has (GST-2268)

The submodule exemption trusted the lane's status, which only compares the submodule's HEAD to the recorded gitlink. Commits on the submodule's own branches or stash live in the lane's git dir and were deleted with exit 0. Uninstall now refuses such a submodule, and matches the gitlink by exact path instead of a prefix or glob pathspec.

Review-Finding: s11/outside-voice#1 reader=outside-voice
Review-Finding: s11/adversarial-native#1 reader=adversarial-native
Review-Finding: p1/specialist-batched#1 reader=testing
Review-Finding: p1/red-team#3 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Explain the submodule rule in the uninstall docs (GST-2268)

docs/uninstall.md now says a submodule inside a clean lane goes only when its commits are on a remote-tracking branch, and the changelog entry drops a mid-sentence colon.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Make every gstack-uninstall keep check run before anything is deleted (GST-2268)

The post-prompt worktree check now runs right after confirmation, before
the install root or anything else is removed, so running uninstall from
inside the gstack install no longer deletes a worktree edited during the
prompt. A .gstack/ holding its own git repository inside the install being
removed now stops the uninstall. A failing git worktree list stops it too,
even with --force-worktrees. The manual cleanup in docs/uninstall.md now
matches worktree paths by prefix, so it never removes a worktree outside
the project whose path merely contains it.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Also stop gstack-uninstall when a .gstack repo sits inside a Cursor, OpenCode, Copilot, Antigravity or Gemini install (GST-2268)

Review-Finding: p1/specialist-batched#1 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Split a run-on sentence in the uninstall release note (GST-2268)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-11-10 07:15:05 +00:00
Garry Tanandcapy-ai[bot] 490bbd5202 Stop gstack-uninstall deleting a repo's committed .gstack files and unsaved worktrees (GST-2197) (#4241)
* Stop gstack-uninstall deleting a repo's committed .gstack files and unsaved worktrees (GST-2197)

--keep-state now keeps the repo's .gstack/ and .gstack-worktrees/. A plain
uninstall keeps git-tracked .gstack files, removes only ignored state, and
stops before removing anything when a worktree has uncommitted changes
unless --force-worktrees is given.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Keep moved, config-hidden and committed worktree work from gstack-uninstall (GST-2197)

The dirty-worktree guard now checks every checkout on disk under .gstack-worktrees/, pins git status flags so config cannot hide new files, refuses a checkout whose commits no branch or tag holds, and keeps .gstack/ when git cannot list its tracked files.

Review-Finding: p1/red-team#2 reader=red-team
Review-Finding: p1/red-team#3 reader=red-team
Review-Finding: p1/specialist-batched#1 reader=security
Review-Finding: p1/red-team#4 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Make gstack-uninstall list worktree checkouts safely and never run their config (GST-2197)

Checkouts are listed NUL-separated and through a symlinked .gstack-worktrees, only worktrees this repository registered are opened with git, and those git calls turn off fsmonitor, hooks and optional index writes.

Review-Finding: v1/adversarial-native#1 reader=adversarial-native
Review-Finding: v1/adversarial-native#5 reader=adversarial-native
Review-Finding: v1/adversarial-native#2 reader=adversarial-native
Review-Loop-Grant: a-f3b2ff8e662e
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Check a .gstack-worktrees that is itself a git repo before gstack-uninstall deletes it (GST-2197)

Review-Finding: fc1/fix-check#1 reader=fix-check
Review-Loop-Grant: a-3b80e7929588
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Document what gstack-uninstall keeps and refuses, and the new --force-worktrees flag (GST-2197)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Stop gstack-uninstall dropping the registration of worktrees it did not delete (GST-2197)

Uninstall ran a repo-wide git worktree prune, which unregistered every worktree whose folder was missing at that moment, including ones outside .gstack-worktrees on an unmounted drive, and left their detached commits unreachable. It now unregisters only the gstack checkouts it just removed.

Review-Finding: p1/red-team#1 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Say that gstack-uninstall does not yet check worktrees nested inside another checkout (GST-2197)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Stop gstack-uninstall deleting checkouts through a .gstack-worktrees symlink that points outside the repo (GST-2197)

Review-Finding: p1/red-team#1 reader=red-team
Review-Finding: p1/red-team#2 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Make gstack-uninstall remove only the link when .gstack-worktrees is a symlink, as before (GST-2197)

Review-Finding: p2/red-team#1 reader=red-team
Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* Drop release-note wording from the gstack-uninstall docs (GST-2197)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-11-09 23:42:47 +00:00
Garry Tanandcapy-ai[bot] b15492a148 v1.92.23.0 feat(harness): opt-in start hook for OpenClaw and Hermes, so agents run a skill's start command (GST-718) (#3786)
* feat(harness): opt-in start hook for OpenClaw and Hermes (GST-718)

gstack-harness register openclaw|hermes --start-hook (or GSTACK_HARNESS_START_HOOK,
or gstack-config harness_start_hook) adds a gstack-start plugin that puts a
one-line note on every user turn to run a gstack skill's start command as
written. Activation goes through each harness's own CLI and is read back;
a run that turns it on and ends inactive exits 4. Off by default.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): retry an inactive start hook without failing refresh, and never re-enable a user disable (GST-718)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): an env or config start-hook opt-in with an unsafe plugin folder turns the hook off instead of failing refresh (GST-718)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): normalize the hook profile, retry an interrupted activation, re-enable after kept edits

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): a non-folder or empty plugin path is never taken as kept edits

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): teardown keeps kept copies and refuses symlinked plugin paths; a failed opt-out can rerun; a saved hook survives OPENCLAW_CONFIG_PATH

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): any saved plugin root is checked for symlinks before an opt-out deletes through it

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): tolerant JSON read-back, short read-back timeout, retry an interrupted explicit activation, symlink checks on every hook opt-out and folder cleanup

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs(harness): start hook unchecked state and uninstall notes

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-11-01 09:52:31 +00:00
Garry Tanandcapy-ai[bot] e4e4e42967 v1.92.22.0 feat: Gemini CLI host, and older installs drop the dev toolchain (#3778)
* setup: explicit contributor marker; move pre-v1.91.231.0 installs to production deps (#3447)

./setup used node_modules/typescript as its contributor signal, so every
install made before v1.91.231.0 (a full bun install) kept ~520 MB of test and
eval packages forever. The contributor is now the clone bin/dev-setup marks
with git config gstack.contributor true, or GSTACK_CONTRIBUTOR=1. An unmarked
clone with the dev toolchain moves node_modules aside under a checkout lock,
installs production-only, verifies every dependency, and drops the backup; any
failure puts the full install back with SETUP_DEPS_MIGRATE_FAILED. CI and a
clone whose git config cannot be read never migrate. Bun cache entries for the
pinned Agent SDK are removed only when no other project links to them.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* hosts: add Gemini CLI as an experimental host (#1247)

./setup --host gemini (alias gemini-cli; --host auto when gemini is on PATH)
installs gstack-<skill> directories into ~/.gemini/skills, which Gemini CLI
0.66.0 lists with gemini skills list, plus the ~/.gemini/skills/gstack runtime
root. Renders go to .gemini-cli/ (gitignored) with name+description
frontmatter whose names match their directories, CLAUDE.md rewritten to
GEMINI.md, and a tool glossary for Gemini CLI's names (and Antigravity's, which
can read the same folder). gstack-uninstall removes only gstack-managed entries
there, so the user's own skills stay. Docs: ADDING_A_HOST.md verified facts,
hosts.md, uninstall.md, README.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix: pre-landing review fixes (leave Bun's shared cache alone)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* changelog: Gemini CLI host and contributor marker (#1247, #3447)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: host lists name Copilot, Antigravity and Gemini CLI

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test: the plan-mode re-render guard count includes the Gemini CLI host

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: troubleshooting's runtime-root guard host list names Gemini CLI

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-11-01 09:35:42 +00:00
Garry Tanandcapy-ai[bot] b61744517c v1.91.334.0 fix(harness): serve OpenClaw's large registered skills as entries pointing at a plain full copy (GST-110) (#3610)
* fix(harness): register OpenClaw skills over its 32 KB whole-read limit as entries pointing at SKILL.full.md (GST-110)

#3555's hybrid layout registered review, investigate, ship and plan-eng-review
as OpenClaw skills; OpenClaw serves a registered SKILL.md whole or refuses it,
at 32,768 bytes when the model's context window is unknown.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs(harness): OpenClaw entry skills and the whole-read limit; CHANGELOG (GST-110)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix: pre-landing review fixes (GST-110)

Entries point at a full SKILL.md in a plain copy of every skill (the router's
gstack/skills/ in hybrid, .gstack-skills/ under --skills all), so autoplan's
../gstack-plan-eng-review/SKILL.md and other sibling paths reach full texts;
test the #3555 upgrade path.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: remove .gstack-skills in the manual OpenClaw uninstall (GST-110)

Step 14.5 audit gst110-docs-1: the gstack* glob never matched the dot folder that --skills all registrations now write.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): OpenClaw entry threshold 32,000 bytes so served skills also fit a 100K-200K-token tool-result budget (GST-110)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* v1.91.334.0 stamp

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-10-30 14:25:11 +00:00
Garry Tanandcapy-ai[bot] 44f19d0106 v1.91.290.0 feat(harness): gstack registers its skills in whatever agent harness it is installed into (GSTA-135) (#3526)
* wip: gstack-harness registrar (GSTA-135)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* wip: harness registration in setup, uninstall, tests

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* wip: harness hosts experimental, troubleshooting codes, kit

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: agent harness registration (GSTA-135)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix: harness hosts run in place from the checkout in --status

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* test: bun-free ledger reader

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: CHANGELOG for harness registration (Unreleased)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): keep edits beside themselves, never follow symlinks, cut only gstack's own context block

Review blockers on #3526: uninstall deleted the edit backups under the state root;
unregister followed a symlinked folder below the destination; a lone or edited
context marker made the cut eat user lines; a re-register adding --context left an
unrecorded block. Also: ledger written before the files, --refresh-registered no
longer re-runs harness hosts through setup (gstack-harness refresh keeps their mode),
bash goes through runBashSync, and setup tests follow openclaw/hermes becoming installs.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix(harness): a ledger with no destination row touches nothing (no hang)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: harness registration in README, install and hosts (document-release audit)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-10-28 04:33:53 +00:00
Garry Tanandcapy-ai[bot] ab9a854be2 v1.91.246.0 docs(readme): cut the README to the first ten minutes; move everything else to docs/ pages verbatim (GSTA-99) (#3428)
* docs(readme): cut the README to the first ten minutes; move everything else to docs/ pages verbatim (GSTA-99)

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix: pre-landing review fixes

Restore the /setup-gbrain code block the README split left open: README.md kept
the ```bash opener and docs/gbrain-quick-tour.md got only the closing fence, so
both pages rendered their body as one code block. Drop the self-referential
'Full reference' link in docs/domain-skills.md and make the troubleshooting
quick-answer link an in-page anchor.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: describe the proactive core-set default in why-gstack

The moved 'Proactive skill suggestions' paragraph now documents the four-skill
core set, gstack-config set proactive core|all|off and the docs/proactive.md
page that #3432 (gsta-99-proactive-core) adds; the CHANGELOG names it as the one
non-verbatim paragraph of the move.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* fix: pre-landing review fixes (round 2)

Promote docs/uninstall.md's moved sections from ### to ## so the page outline
matches the other pages the README split created; no inbound link targets
those anchors.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: documentation audit fixes

The README Docs table no longer claims docs/standalone-binaries.md covers
every gstack-* CLI (it documents the standalone ones), and
docs/ADDING_A_HOST.md step 6 points at the README host-matrix row and
docs/hosts.md, where the per-host notes now live.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: disclose the update check fully and list every setup hook

The README privacy summary now names both GitHub requests the update check
makes and the version/OS report it adds with telemetry on, as the full text
it replaced did. docs/install.md's hooks section lists the team-mode
SessionStart hook and the AskUserQuestion PreToolUse/PostToolUse hooks beside
the timeline Stop hook.

* docs: name the update check's conditional compare-API request

bin/gstack-update-check also GETs api.github.com's compare endpoint when the
remote VERSION equals the local one but main has moved; the README privacy
summary and docs/privacy.md now list it beside the ls-remote and VERSION
fetch (documentation audit).

* fix: pre-landing review fixes (merge resolution)

docs/install.md:21 carried main's stable-channel sentence twice after the
merge, the second copy with its backticked spans stripped; the paragraph
again matches README.md:41 verbatim with the relative link.

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

* docs: say outside reviews are on by default in the README privacy summary

The one-paragraph summary dropped the default-on disclosure the full text
kept; mirror docs/privacy.md's parenthetical so the top-level page says which
path sends code off-machine without an opt-in.

* v1.91.246.0 stamp

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>

---------

Co-authored-by: capy-ai[bot] <230910855+capy-ai[bot]@users.noreply.github.com>
2026-10-26 13:22:30 +00:00