Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add daeryundf2-prog/LAZYANTIGRAVITY --skill mass-ulwgit clone --depth 1 https://github.com/daeryundf2-prog/LAZYANTIGRAVITYWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/daeryundf2-prog/lazyantigravity/mass-ulw)<a href="https://agentmods.dev/skills/daeryundf2-prog/lazyantigravity/mass-ulw"><img src="https://agentmods.dev/badge/skills/daeryundf2-prog/lazyantigravity/mass-ulw/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/daeryundf2-prog/lazyantigravity/mass-ulw"><img src="https://agentmods.dev/badge/skills/daeryundf2-prog/lazyantigravity/mass-ulw.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00067 | $0.01222 |
| Opus 5.5 | $0.00027 | $0.00489 |
| Sonnet 5.5 | $0.00013 | $0.00244 |
| Haiku 4.5 | $0.00007 | $0.00122 |
Grade A, and why
mass-ulw scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
mass-ulw
Use this skill when the work splits into many pieces and some pieces depend on others. If every piece is independent, a single invoke_subagent call with several entries is enough; you do not need this skill.
Antigravity has no DAG engine. You are the scheduler: you hold the graph, launch one wave at a time with a single invoke_subagent call, and start the next wave only after the current one reports back. Read references/planning.md in this skill's directory before defining the first wave.
1. Lock the topology
Before writing any node:
- List the 1 to 6 top-level components that can each succeed or fail on their own. Every node belongs to exactly one.
- For each component, ask how to break it into small pieces that can run at the same time. Keep a piece whole only when splitting it would put two writers on one file, cut a single judgment call in half, or cost more in briefing than in doing.
- Draw the dependency edges. An edge means "B needs A's finished result". Never add an edge just to pass along a fact you already know; paste the fact into B's prompt instead.
- Self-check: no cycles, every edge points at a real node, every wave has a runnable node, no two nodes in the same wave write the same file.
Write the plan out before launching: waves, node ids, model per node with a one-line reason for anything above flash, and the verification wave.
2. Route each node
Start every node at the cheapest lane and climb only with a stated reason.
| Lane | Model |
Use for |
|---|---|---|
| lookup | flash_lite |
file listing, single grep, lint read |
| default | flash |
mechanical or single-area code, search, drafting |
| hard | pro |
cross-module reasoning, security, architecture, adversarial review |
| verdict | inherit |
at most one node per graph, only on a Claude session, for the final call |
Agent types: explore for in-workspace search, librarian for upstream code and docs, research for read-only broad surveys, self for anything that writes. A wave where every node is pro is a routing mistake.
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 11d ago First seen · 75 lines · 67 tokens per session scan A e39de229fd73
mass-ulw is a skill published in the GitHub repository daeryundf2-prog/LAZYANTIGRAVITY (14 stars, last pushed 3d ago), licensed MIT. It adds 67 tokens to every session and 1,222 once invoked, about $0.0003 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-05.
Other skills, from other repositories
top-down-typescript
TypeScript and JavaScript coding style conventions, centered on top-down narrative ordering (caller first, helpers below) and functions over classes. Read before writing or reviewing TypeScript/JavaScript code, including code inside a spec or a plan.
loom-typescript
TypeScript language expertise for type-safe, production-quality code.
ai-provider-anthropic-sdk
Official Anthropic SDK patterns for TypeScript/Node.js — client setup, Messages API, streaming, tool use, vision, extended thinking, structured outputs, prompt caching, batch API, and production best practices.
ai-infrastructure-ollama
Local LLM inference with the Ollama JavaScript client -- chat, streaming, tool calling, vision, embeddings, structured output, model management, and OpenAI-compatible endpoint.
ai-infrastructure-replicate
Replicate SDK patterns for TypeScript/Node.js -- client setup, predictions, streaming, webhooks, file handling, model versioning, deployments, and training.
ai-infrastructure-together-ai
Together AI SDK patterns for TypeScript — client setup, chat completions, streaming, structured output, function calling, embeddings, image generation, fine-tuning, and OpenAI-compatible endpoints.