New: 97,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

agystack Runtime Setup

by jtaroreh in jtaroreh/agystack

Configures agystack's model tiers per role and its execution runtime, choosing between local subagents and Cloud Run jobs for large parallel swarms.

MITAuto-check passedAgent Workflows

Install agystack Runtime Setup

skills CLI
$ npx skills add jtaroreh/agystack --skill setup-agystack -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install jtaroreh/agystack setup-agystack --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/jtaroreh/agystack.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/setup-agystack .claude/skills/setup-agystack && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
setup-agystack
GitHub stars
103
Token cost
~1.7k tokens
SKILL.md length
728 words
Files
2 (incl. scripts)
Skills in repo
6
Repo updated
First seen
Licence
MIT

At a glance

Configures agystack's model tiers per role and its execution runtime, choosing between local subagents and Cloud Run jobs for large parallel swarms.

  • Works in 7 steps: Select execution runtime → Detect available models → Load current state → …
  • Setting up agystack for the first time
  • SKILL.md covers Prerequisites and Steps
  • Runs Python scripts from its folder; calls python3; needs GEMINI_API_KEY

What it does

The skill writes model tiers to agystack-models.md and runtime settings to agystack-runtime.json under the Gemini config plugins folder, or the workspace .agents/plugins folder. Prerequisites are bun for PR babysitting and orchestration, with Node.js not supported, gh for PR automation, gt recommended for stacked PRs, and optional Google Cloud Python libraries, all checked by running scripts/setup_runtime.py with --doctor.

Step one is choosing a runtime. The local default uses native subagent invocation in Antigravity for up to 8 concurrent workers with no cloud setup, while Cloud Run Jobs supports 10 to 100+ parallel workers in isolated containers and only starts when you explicitly trigger /swarm. The agent must explain that Cloud Run is an on-demand batch runner that costs nothing when idle and that daily work such as pair programming, bug fixes and reviews always stays local. The excerpt is cut off before the Cloud Run steps.

When your agent uses it

  • Setting up agystack for the first time
  • Assigning model tiers to agystack roles
  • Switching a swarm from local workers to Cloud Run Jobs
  • Running the doctor check to verify agystack dependencies

Example prompts

  • “Set up agystack with the local runtime and run the doctor check.”
  • “Switch agystack to Cloud Run so I can swarm a refactor across many workers.”
  • “Change the model tier for the reviewer role in agystack.”

Requirements

  • bun, for watch-pr and orch
  • GitHub CLI (gh) for PR automation
  • Google Antigravity with the agystack plugin

Workflow steps

7 steps, taken from the step headings in SKILL.md.

  1. Select execution runtime
  2. Detect available models
  3. Load current state
  4. Map and confirm
  5. Write the model rule
  6. Confirm
  7. Offer a verification skill (optional)

What it can do on your machine

Read from SKILL.md and the folder at commit d1a0466. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Ships 1 file in scripts/ (Python), which the agent can run.

    Shell commands in SKILL.md call:

    • python3

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names these keys or tokens, usually read from environment variables:

    • GEMINI_API_KEY

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

agystack Runtime Setup loads about 1.7k tokens when it runs. Until then it costs about 44 tokens; SKILL.md has 728 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~44
When it runs · the whole SKILL.md, loaded when a task matches
~1.7k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.

SKILL.md

The full file from jtaroreh/agystack at commit d1a0466, republished under its MIT licence (© jtaroreh). 728 words, ~1,729 tokens.

Download SKILL.mdSave it as .claude/skills/setup-agystack/SKILL.md (or your agent's skills folder). This skill also uses 1 other file; get the full folder from GitHub.
name
setup-agystack
description
Configure models and execution runtime for agystack. Configures Antigravity model tiers per role and Cloud Run runtime settings for up to 100+ parallel workers.

Setup agystack

Configure model tiers in ~/.gemini/config/plugins/agystack/rules/agystack-models.md and execution runtime in ~/.gemini/config/plugins/agystack/agystack-runtime.json (or workspace .agents/plugins/agystack/).

Prerequisites

  • bun (v1.0+) is mandatory for PR babysitting (watch-pr) and multi-agent orchestration (orch). Node.js is not supported.
  • gh (GitHub CLI) is required for PR automation and preflight checks.
  • gt (Graphite CLI) is recommended for stacked PRs.
  • google-cloud-storage, google-genai, and google-cloud-run are optional Python libraries for GCS artifact storage, preflight checks, and Cloud Run job monitoring.

During setup, verify dependencies automatically via --doctor:

bash
python3 "$(find ~/.gemini/config/plugins/agystack .agents/plugins/agystack skills/setup-agystack -name "setup_runtime.py" 2>/dev/null | head -1)" --doctor

Steps

1. Select execution runtime

Choose the execution runtime for parallel swarms:

  • Local Runtime (Default): Runs via native invoke_subagent in Antigravity for up to 8 concurrent workers. Zero cloud setup required.
  • Cloud Run Runtime: Runs via Google Cloud Run Jobs for 10 to 100+ parallel workers in isolated container instances. Cloud Run is strictly an on-demand batch runner. It only spins up containers when you explicitly trigger /swarm (or ask to swarm a task across many parallel workers). It does not run continuously and is never an always-on server. It costs $0 when idle. Daily tasks (pair programming, routine edits, bug fixes, refactoring, code reviews via /interrogate, and local subagents) always run locally on your machine.

CRITICAL INSTRUCTION FOR AI AGENT WHEN PRESENTING RUNTIME CHOICE: When presenting the runtime choice to the user, explicitly explain the cost and execution model before asking them to choose:

  1. Explain that Cloud Run is strictly an on-demand batch runner. It only spins up containers when the user explicitly triggers /swarm (or asks to swarm a task across many parallel workers).
  2. Clarify that Cloud Run does NOT run continuously and is never an always-on server.
  3. State that Cloud Run costs $0 when idle.
  4. Reassure the user that daily tasks (pair programming, routine edits, bug fixes, refactoring, code reviews via /interrogate, and local subagents) ALWAYS run locally on their machine.

If Cloud Run is selected: CRITICAL INSTRUCTION FOR AI AGENT FOR PROVISIONING: NEVER print manual bash commands with placeholders (like <your-gcp-project-id>) for the user to run. You MUST directly execute the setup commands yourself using python3 "$(find ~/.gemini/config/plugins/agystack .agents/plugins/agystack skills/setup-agystack -name "setup_runtime.py" 2>/dev/null | head -1)" right here in the chat environment.

  1. Verify Quotas and Permissions:
    • Google AI Studio API Key: If using GEMINI_API_KEY, verify that paid billing (Pay-as-you-go / Tier 1+) is enabled on the AI Studio project. Free-tier API keys (capped at 5 requests per minute) are strictly prohibited for swarms because parallel workers will hit immediate rate limits.
    • Vertex AI Mode: If using Vertex AI mode, verify that the GCP project has the Vertex AI API enabled (aiplatform.googleapis.com) and that the active user or service account has the Vertex AI User role (roles/aiplatform.user).
    • Model Availability: Gemini 3 series models (gemini-3.8-flash) require global routing (aiplatform.googleapis.com with locations/global). Regional endpoints return HTTP 404 for Gemini 3.x.
  2. Check Requirements: Run python3 skills/setup-agystack/scripts/setup_runtime.py --check to verify gcloud is available and authenticated.
  3. Select Project:
    • Query available projects by running python3 skills/setup-agystack/scripts/setup_runtime.py --list-projects.
    • Ask the user which project they want to use, or if they want you to create a new one.
  4. Provisioning: Once a project ID is known, run the full provisioner (do this yourself, do not ask the user to do it!): python3 skills/setup-agystack/scripts/setup_runtime.py --project <PROJECT_ID> --auto-provision --scripts-dir skills/swarm/scripts (or use --create-project <PROJECT_ID> instead of --project if creating a new one).
Show full SKILL.md (181 more words)Show less

Do not leave the user to do the work. Complete the deployment end-to-end for them. Ensure GEMINI_API_KEY is exported in the user's environment with paid tier enabled (Pay-as-you-go), or Vertex AI permissions and global endpoint access are verified.

2. Detect available models

Enumerate the model tiers you can pass to invoke_subagent:

  • pro: High-capability tier (maximum reasoning budget for complex code, architecture, and hard tasks)
  • flash: Balanced fast tier (fast execution for exploration and standard generation)
  • flash_lite: Lightweight tier (minimal latency for quick lookups)
  • inherit (or auto): Inherit parent chat model
3. Load current state

If ~/.gemini/config/plugins/agystack/rules/agystack-models.md exists, read its current role assignments. Otherwise start from skill defaults.

4. Map and confirm

Show every role with its model tier and confirm:

  • Single roles: feature, refactoring, bug-fix, perf-issue, hillclimb, swarm workers
  • Panel roles: arena runners, architect runners, interrogate reviewers
5. Write the model rule

Write to .agents/plugins/agystack/rules/agystack-models.md if installed workspace-locally, otherwise ~/.gemini/config/plugins/agystack/rules/agystack-models.md:

# agystack model configuration. One line per role. Delete a line to fall back to the skill default.
# Antigravity model tiers for invoke_subagent:
# - pro        (High-capability tier: deep reasoning, large refactors, complex design)
# - flash      (Balanced fast tier: exploration, reading, standard code generation)
# - flash_lite (Lightweight tier: fast mechanical lookups)
# - inherit    (Runs on the active parent chat session model)

feature, refactoring: pro
bug-fix: pro
perf-issue: pro
hillclimb: pro
judgment and prose: pro
hardest tasks: pro
how explorer: flash
how explainer: pro
why investigators: flash
why synthesizer: pro
reflect tooling: pro
reflect judgment, divergent, synthesizer: pro
arena runners: pro, flash, inherit
arena cross-judge pool: pro, flash, inherit
swarm workers: flash
architect runners: pro, flash, inherit
interrogate reviewers: pro, flash, inherit
6. Confirm

Confirm that the model rule and runtime settings are active for new sessions.

7. Offer a verification skill (optional)

If the project lacks an end-to-end verification harness, offer /create-verification-skill.

© jtaroreh, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 1 other file (scripts) in skills/setup-agystack of jtaroreh/agystack.

  • SKILL.md
  • scripts/setup_runtime.py

Open the folder on GitHubat commit d1a0466

Compare with similar skills

agystack Runtime Setup next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

agystack Runtime Setup compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
agystack Runtime Setup this skilljtaroreh/agystack103—~1.7kAutomated safety check: PassMIT
Subagent-Driven DevelopmentHoangNguyen0403/agent-skills-standard570—~852Automated safety check: PassMIT
Argentos Dev OpsArgentAIOS/argentos-core126—~1kAutomated safety check: PassCustom licence
Paseo Advisor Second Opiniongetpaseo/paseo20k1 repos~756Automated safety check: PassCustom licence
O2 Review Loopopenobserve/openobserve22k—~3.7kAutomated safety check: PassAGPL-3.0
Harness Agent Team Designerrevfactory/harness9.1k—~4.5kAutomated safety check: PassApache-2.0

Similar skills

  • Subagent-Driven Development

    HoangNguyen0403/agent-skills-standard

    Runs a multi-task implementation plan by sending each task to a fresh implementer subagent, reviewing it independently, then reviewing the whole branch.

    570 GitHub stars~852 tokensUpdated today
    Agent WorkflowsAuto-check passed

  • Argentos Dev Ops

    ArgentAIOS/argentos-core

    ArgentOS development operations workflow — Linear tracking, CodeRabbit reviews, Blacksmith CI, GitHub branch protection, and inter-agent handoff protocol.

    126 GitHub stars~1k tokensUpdated 3 mo ago
    Agent WorkflowsAuto-check passed

  • Paseo Advisor Second Opinion

    getpaseo/paseo

    Launches one separate agent through Paseo to give a second opinion on the current task, with a self-contained briefing and no permission to edit files.

    20k GitHub starsUsed in 1 repo~756 tokens
    Agent WorkflowsAuto-check passed

  • O2 Review Loop

    openobserve/openobserve

    Splits a change into planner, coder and independent reviewer roles: you confirm a spec, a subagent implements it, and a separate reviewer checks each round's local WIP commit.

    22k GitHub stars~3.7k tokensUpdated today
    Agent WorkflowsAuto-check passed

  • Harness Agent Team Designer

    revfactory/harness

    Designs a project-specific agent harness: defines specialist agents, writes the skills they follow, picks an execution mode and model for each, and keeps the setup maintained.

    9.1k GitHub stars~4.5k tokensUpdated 9 days ago
    Agent WorkflowsAuto-check passed

  • ClawTeam Multi-Agent Swarm

    win4r/ClawTeam-OpenClaw

    Launches a swarm of specialist Hermes agents in git-worktree-isolated tmux windows with a kanban board and file-based inboxes, using built-in templates like hedge-fund and code-review.

    1.5k GitHub starsUsed in 1 repo~2.9k tokens
    Agent WorkflowsAuto-check passed

More from jtaroreh/agystack

  • Loop Until Verified

    jtaroreh/agystack

    Runs an iterative verification loop on Google Antigravity: make one edit, run a test command, and reschedule until the command passes or the iteration budget runs out.

    103 GitHub stars~607 tokensUpdated 5 days ago
    Auto-check passed

  • Deslop

    jtaroreh/agystack

    Clean AI slop and defensiveness out of code and diffs before commit.

    103 GitHub stars~890 tokensUpdated 5 days ago
    Auto-check passed

  • agystack Engineering Framework

    jtaroreh/agystack

    agystack's agentic engineering framework and rigor mode. Runs poteto-mode playbooks, isolated subagents, and cloud swarms. Use for /agystack, agystack, or…

    103 GitHub stars~322 tokensUpdated 5 days ago
    Auto-check passed

  • Principle Attack The Premise

    jtaroreh/agystack

    Apply when handed a problem, constraint, or design you are tempted to accept as given.

    103 GitHub stars~489 tokensUpdated 5 days ago
    Auto-check passed

Works with

Categories

Questions about agystack Runtime Setup

What does agystack Runtime Setup do?

Configures agystack's model tiers per role and its execution runtime, choosing between local subagents and Cloud Run jobs for large parallel swarms. agents/plugins folder.py with --doctor.

When should I use agystack Runtime Setup?

agystack Runtime Setup fits situations like: setting up agystack for the first time; assigning model tiers to agystack roles; switching a swarm from local workers to Cloud Run Jobs; running the doctor check to verify agystack dependencies.

How do I install agystack Runtime Setup in Claude Code?

Run `npx skills add jtaroreh/agystack --skill setup-agystack -a claude-code`. Or copy the skill folder (skills/setup-agystack in jtaroreh/agystack) into .claude/skills/setup-agystack in your project. Claude Code loads it when a task matches its description.

How do I install agystack Runtime Setup in Codex?

Run `npx skills add jtaroreh/agystack --skill setup-agystack -a codex`. Or copy the skill folder (skills/setup-agystack in jtaroreh/agystack) into .agents/skills/setup-agystack in your project. Codex loads it when a task matches its description.

Can I use agystack Runtime Setup in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add jtaroreh/agystack --skill setup-agystack -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/setup-agystack, .gemini/skills/setup-agystack, .github/skills/setup-agystack and .opencode/skills/setup-agystack in your project.

What does agystack Runtime Setup need to run?

Going by SKILL.md and its folder, agystack Runtime Setup needs Python for the scripts in its folder, the command-line tools its instructions call (python3) and credentials named GEMINI_API_KEY. Our summary lists: bun, for watch-pr and orch; GitHub CLI (gh) for PR automation; Google Antigravity with the agystack plugin.

Does agystack Runtime Setup access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is agystack Runtime Setup safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.

What licence does agystack Runtime Setup use?

agystack Runtime Setup is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does agystack Runtime Setup use?

About 1.7k tokens (SKILL.md is roughly 6.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to agystack Runtime Setup?

Skills that share tags, products or a category with agystack Runtime Setup: Subagent-Driven Development (HoangNguyen0403/agent-skills-standard, 570 stars), Argentos Dev Ops (ArgentAIOS/argentos-core, 126 stars), Paseo Advisor Second Opinion (getpaseo/paseo, 20k stars) and O2 Review Loop (openobserve/openobserve, 22k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains agystack Runtime Setup?

jtaroreh (a GitHub user) maintains it in jtaroreh/agystack, which has 103 GitHub stars. The repository holds 6 skills in this directory. The repository was last updated on October 1, 2026.

Source: jtaroreh/agystack on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.