explainx / blog
gstack packages YC-style slash skills—office hours, plan reviews, /review, /qa in a real browser, /cso, /ship—plus power tools, OpenClaw integration, and optional CLIs. Here is a detailed map of the repo, multi-host install, and how it fits explainx.ai’s view of agent skills.
Jun 28, 2026
Most agent demos work once and fail in production. This guide covers the architecture of a real ai agent loop — triggers, state, executors, terminators, retry logic, checkpoints, and the human handoff decision framework — with working code patterns for all three.
May 12, 2026
Goal mode is an autonomous agent loop where you define a completion condition—like 'all tests passing' or 'vulnerability patched'—and the agent iterates continuously (plan → act → test → review) until the goal is met or your budget is exhausted. Unlike traditional prompt-response cycles, goal mode enables true hands-off automation for complex, multi-step tasks.
May 6, 2026
When agents need real browsers—not just fetch—Browserbase wires stealth sessions, captchas, traces, and cookie sync into skill packs. Here is the upstream inventory and install paths.
gstack is Garry Tan’s public release of a large, Claude Code-centric, multi-host skill stack: on the order of 23 specialist slash workflows, eight “power” tools, MIT license, no paywall. The README leads with a reproducible take on “logical” change and line counts (On the LOC Controversy) and with GitHub contribution density (2026 vs 2013) on the author’s own repositories—then argues for gstack as the sprint process he uses in the open.
This explainx.ai note is a map, not a replacement for upstream docs. github.com/garrytan/gstack is canonical for install commands, slash names, and per-release features (for example the v0.19 CLIs mentioned in the README). Browse gstack on explainx.ai: explainx.ai/skills/gstack (pre-filtered skills registry view).
Turn a long chat into a repeatable sprint where the output of one slash skill feeds the next: think → plan → build → review → test → ship → reflect.
Upstream frames that as a process, not a grab bag of prompts. It overlaps the “skills compound if you chain them” idea in our agent skills guide, at YC-style cadence.
| Archetype | Examples (see latest README) |
|---|---|
| Product and planning | /office-hours, /plan-ceo-review, /plan-eng-review, /plan-design-review, /plan-devex-review, /autoplan |
| Design | /design-consultation, /design-shotgun → /design-html (Pretext-style responsive layout, framework detection—see docs) |
| Code quality | /review (Claude), /codex (OpenAI Codex CLI “second opinion” with multiple modes), /investigate |
| Runtime truth | /qa / /qa-only, /browse, /open-gstack-browser, /setup-browser-cookies |
| Security | /cso (OWASP + STRIDE, per upstream copy) |
| Release | /ship, /land-and-deploy, /canary, /benchmark, /document-release |
| Meta | /retro, /learn, optional checkpoint mode with /context-restore, /gstack-upgrade |
| Safety | /careful, /freeze, /guard, /unfreeze |
| Multi-agent | /pair-agent (shared GStack Browser with tab isolation, optional remote use—see docs) |
Philosophy per skill: docs/skills.md. Builder framing: ETHOS.md.
~/.claude/skills/gstack, run ./setup, add a CLAUDE.md gstack block that (a) routes web work through /browse and (b) lists slash commands. The long canonical list is only in the upstream README; do not copy stale lists from blogs.gstack-team-init in required or optional mode commits .claude/ metadata so peers auto-install; setup --team plus silent hourly update checks../setup --host <name> to install under paths such as ~/.codex/skills/, ~/.cursor/skills/, and others listed in the README. Extending a host: docs/ADDING_A_HOST.md.Requirements (per README): Claude Code, Git, Bun 1+; Windows also Node (Playwright + Bun pipe issue—upstream documents fallback to Node).
gstack-taste-update to persist what you keep vs reject.$B handoff for MFA or CAPTCHAs, and a substantial prompt-injection story (classifiers, canary tokens, optional heavier ensemble, kill-switch env vars).Before you aim this at untrusted pages or a compliance boundary, read ARCHITECTURE.md and BROWSER.md. This post is not a security audit.
gstack-model-benchmark (Claude, GPT, Gemini; --dry-run to validate), gstack-taste-update (persist design shotgun taste)./codex is a cross-model second opinion on the same diffs as /review, with overlap/unique-finding analysis when both have run.[gstack-context]; /ship is documented to squash or filter WIP commits in a bisect-safe way (see the README).supabase/migrations/. Local gstack-analytics on JSONL./office-hours on a real user pain, not a feature wish list./review on a real branch./qa on a staging URL, then /ship.Hiring (upstream README): Y Combinator — software (not affiliated with explainx.ai; listed for completeness).
The GitHub default branch README and CHANGELOG are the canonical feature and version sources; star counts and social screenshots go stale in hours.