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 agentmods add skills/echoo19/hearth/hearth-designnpx skills add echoo19/hearth --skill hearth-designgit clone --depth 1 https://github.com/echoo19/hearthWrote 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/echoo19/hearth/hearth-design)<a href="https://agentmods.dev/skills/echoo19/hearth/hearth-design"><img src="https://agentmods.dev/badge/skills/echoo19/hearth/hearth-design.svg" alt="Measured on agentmods" 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 | $0.00109 | $0.02851 |
| Opus 5 | $0.00055 | $0.01425 |
| Sonnet 5 | $0.00022 | $0.00570 |
| Haiku 4.5 | $0.00011 | $0.00285 |
Grade A, and why
hearth-design 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 6d 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 — 230 lines — stays where its author put it; the contents beside it link to each section on GitHub.
The shape of a complete game
The sibling skills make a game work: structure is hearth-build, behavior is
hearth-code, art is hearth-art, moment-to-moment polish is hearth-feel.
This skill is the macro layer — how much game to build, in what order the
player meets it, and what lets them feel finished instead of abandoned. The
most common failure of an agent-built game is not a bug. It is a game that
stops rather than ends: one scene, one mechanic, a death loop, no win state.
Load this skill when planning, ideally before hearth init. Scope decisions
cost nothing at init and a rebuild once five scenes exist.
Scope to session length
Pick a target session length first, then spend a budget:
verbs × levels × enemies. Every verb (a distinct player action) multiplies
work across the other skills — an input action (hearth set-input), animation
states, a juice stack, and playtests of its own. Levels and enemy types
multiply the same way.
| Target session | Verbs | Play scenes | Enemy/hazard types |
|---|---|---|---|
| ~5 minutes (arcade / jam scope) | 1–2 | 1–3, plus menu + ending | 2–3 |
| ~20 minutes (small complete game) | 2–3 | 4–6, plus menu + ending | 4–5, the last reserved for the finale |
Depth over breadth — cut a mechanic before shipping two shallow ones. A
verb earns its slot by being taught, developed, and present in the final
challenge. If it won't appear in the climax, cut it and deepen what remains
(more enemy interactions with the existing verb beats a second half-tuned verb).
The genre templates (hearth init --template platformer|topdown|arcade) start
at one or two verbs on purpose: get that verb green in a playtest and juiced
per hearth-feel before even considering another.
Scope is inspectable, so check it instead of estimating: hearth inspect project --json lists every scene, and hearth recall holds the scope you
committed to. Record the budget at the start (hearth remember "Scope: 2 verbs (move+shoot), 4 levels, 3 enemy types" --section decision) so later sessions
inherit the boundary instead of quietly growing it.
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.
- 6d ago First seen · 230 lines · 109 tokens per session scan A 479d3a2a3750
hearth-design is a skill published in the GitHub repository echoo19/hearth (62 stars, last pushed 1mo ago), licensed MIT. It adds 109 tokens to every session and 2,851 once invoked, about $0.0005 per session on Opus 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 2026-08-30.
Other skills, from other repositories
self-evolve
Capture reusable patterns from a finished project and lift them into framework-level priors (contracts, modules, skeletons) that future projects inherit. Run only when the user explicitly requests self-evolution; the orchestrator executes the workflow.
artist-self-evolve
Distill stable art-generation patterns from a completed project, so future projects produce comparable assets without re-discovering the prompts. Lead-dispatched only — orchestrator invokes this skill from its self-evolve flow with a game-slug message; do not self-trigger.
vibegame-build
Run VibeGame's standard end-to-end game development workflow with reviewer gates. Use when the user wants to create a game from zero or evolve an existing game across multiple stages.
vibegame-start
Resume a VibeGame orchestrator session after vibegame start. Use at the beginning of a Claude or Codex session to inspect team runtime state, repair missing persistent members, load goal and GDD context, inspect tasks, and ask the user what to do next.
design-npc
Use when the user wants to design an enemy, NPC, boss, companion, civilian, or wave-spawned mob's behavior. Walks perception, personality knobs, intent layer, action state machine, telegraphs, defeat handling, and group emergence — outputs a state-machine GDScript stub plus the recommended node tree. Trigger on…
fps-controller
Use when building an FPS, first-person movement, or a 3D character that walks, jumps, and reacts to external forces — production-quality first-person controller with WASD, mouse look, jump, coyote time, jump buffering, air control, and external-velocity handling. Trigger on "FPS", "first-person", "WASD", "character…