skills.rest
aicodepro
aicodeproOFFICIAL
Explore all skills by this author →
GitHub

gstack

Provides structured AI agent workflow skills and a persistent Chromium browser daemon for software development tasks.

Updated Jun 22, 2026
Search Tags:#code review#browser automation#ai agent#playwright#release workflow#qa testing#chromium daemon
One-click install
npx skills add https://github.com/aicodepro/ai-agent-nexi --skill gstack-aicodepro
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: gstack
Source: https://github.com/aicodepro/ai-agent-nexi/tree/main/agent/skills/gstack
Command: npx skills add https://github.com/aicodepro/ai-agent-nexi --skill gstack-aicodepro

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires playwright, bun, @huggingface/transformers, and includes scripts (resource) and references (resource) components.

What problem does it solve? AI coding agents lack structured workflows for planning, reviewing, testing, and shipping software, and browser automation for agents is slow and stateless. gstack supplies a collection of specialist SKILL.md workflows (CEO review, eng review, QA, ship, debug) plus a persistent headless Chromium daemon that executes browser commands in ~100-200ms with persistent cookies, tabs, and login sessions. ## Core Features & Use Cases - Workflow skills for the full dev lifecycle: plan-mode reviews (/office-hours, /plan-ceo-review, /plan-eng-review), implementation review (/review, /investigate, /qa), and release automation (/ship, /land-and-deploy, /canary). - Persistent browser daemon: a compiled Bun CLI drives headless or headed Chromium over localhost HTTP with ref-based element selection (@e refs from accessibility snapshots), cookie import, screenshots, PDF export, and ~70 commands. - Codifiable browser automation: /scrape prototypes a data-extraction flow once, /skillify converts it into a deterministic Playwright script, and subsequent runs complete in ~200ms. - Use Case: A developer asks the agent to review a PR, run QA against a staging URL in a real browser, fix the found bugs, and ship the branch with a version-bumped PR — all through coordinated skills. ## Quick Start Ask the agent to run /office-hours to reframe your product idea, or run /qa to open a real browser, find bugs in your app, and fix them.

Frequently Asked Questions about gstack

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I give an AI coding agent structured review workflows?▼

Install gstack's SKILL.md collection, which defines specialist roles like /plan-ceo-review, /plan-eng-review, /review, and /ship. Each skill encodes a multi-step methodology the agent follows when invoked by name.

How do I automate browser tasks from an AI agent without slow restarts?▼

gstack runs a persistent Chromium daemon that the CLI talks to over localhost HTTP. The first call starts the browser in about 3 seconds; every subsequent command completes in roughly 100-200ms with cookies and tabs preserved.

What is the difference between /scrape and /skillify in gstack?▼

/scrape prototypes a data-extraction flow by driving the page once with browser commands. /skillify codifies that successful flow into a deterministic Playwright script, so the next /scrape on the same intent runs in about 200ms.

Does gstack work on Windows?▼

A curated Windows-safe test subset runs on windows-latest CI, and state paths resolve through CLAUDE_PLUGIN_DATA or GSTACK_HOME for cross-platform installs. The setup script currently requires Git Bash or MSYS; native PowerShell support is planned.

How does gstack defend against prompt injection from web pages?▼

It layers defenses: content sanitization and datamarking on page output, a local BERT-small ONNX classifier scanning tool outputs, a transcript-level classifier, a canary token that blocks sessions if leaked, and an ensemble combiner requiring two classifiers to agree before blocking.

Why do gstack browser refs become stale after page changes?▼

Refs are cleared on navigation and SPAs can mutate the DOM without navigating. The resolveRef check detects missing elements and fails fast with a message telling the agent to re-run snapshot for fresh refs.

Related Skills

Explore More Software Engineering Skills→
yogogml0035-cpuyogogml0035-cpu

review-push

Automate code review, testing, and deployment checks for Git branches.

Community
Advanced
ShmurdocShmurdoc

land-and-deploy

Merge pull requests, monitor CI pipelines, and verify production canary health.

Community
Advanced
AutomatedMarketerAutomatedMarketer

security-reviewer

Scan application code, configuration, and dependencies for security vulnerabilities.

Community
Advanced
verrio1verrio1

Browser

Automate browser tasks with Playwright and capture console, network, and page errors.

Community
Intermediate