Borrowing it
Nothing to install: this file belongs to kamwoh/yume. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/kamwoh/yume/dev/.claude/skills/yume-map-author/SKILL.mdgit clone --depth 1 https://github.com/kamwoh/yumeWrote 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/kamwoh/yume/yume-map-author)<a href="https://agentmods.dev/skills/kamwoh/yume/yume-map-author"><img src="https://agentmods.dev/badge/skills/kamwoh/yume/yume-map-author/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/kamwoh/yume/yume-map-author"><img src="https://agentmods.dev/badge/skills/kamwoh/yume/yume-map-author.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 | $0.00050 | $0.04009 |
| Opus 5.5 | $0.00020 | $0.01604 |
| Sonnet 5.5 | $0.00010 | $0.00802 |
| Haiku 4.5 | $0.00005 | $0.00401 |
Grade A, and why
yume-map-author 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 12d 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 — 392 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/yume-map-author
You are the level / map author for Yume. You read a semantic
top-down map image and produce a level's entities.json — meaning
the in-game level matches the map (modulo procedural-scatter
variance for zones).
Sibling of yume-hud-author and yume-screen-author. Same harness
pattern; the surface this time is a 2D spatial world, not a 16:9
viewport. Deterministic surface lives at
tools/visual_layout/wireframe_to_map.py.
⚠ CURRENT PIPELINE (2026-05-27). The LLM-parses-the-map-PNG flow below (wireframe_to_map) places entities by reading the map via vision against a per-game entity catalog. It has been largely SUPERSEDED for FULL-WORLD generation by
tools/visual_layout/compose_world.py, which extracts placements DETERMINISTICALLY from the semantic map viadata/lib/extraction_strategies.json+lib_extract_dispatch(connected components, PCA fit-to-mask, variant buckets, kit meshes), and emits entity defs + biome ground + water + roads. A separatecompose_shell.pyadds camera/player/input/lighting. This skill still applies when authoring a level by hand-placing catalog entities against a wireframe (the fit-fit LLM-parser path), but text-to-WORLD generation now goes through compose_world/compose_shell (the active 3D map pipeline; see.claude/rules/pipeline-stability.md).
Why this skill exists
The earlier compose_map.py chain used opencv k-means + skimage
skeletonization to parse semantic maps. Same fragility as the
CV-based HUD pipeline: borderline-cropped color regions, soft
JPEG edges, ambiguous zone/anchor calls.
Putting Claude at the parsing step gives:
- Reliable region identification via vision (not k-means)
- Semantic def_id picking ("this blob looks like a fire pit because it's small and at center" — informed by tags + asset_resolution hints)
- Better anchor-vs-zone judgment (a small distinct region is an anchor; a large irregular region is a zone)
Fit-fit: what's drawn on the map is what's placed. No clever "densify the forest by 2x" or "shift the camp 10m east." The map is the spec.
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.
- 12d ago Changed · +12 lines · -118 tokens per session 6d4cc38c4865
- 1mo ago First seen · 380 lines · 168 tokens per session scan A a9753a01ab26
yume-map-author is a skill published in the GitHub repository kamwoh/yume (25 stars, last pushed 10d ago), licensed MIT. It adds 50 tokens to every session and 4,009 once invoked, about $0.0002 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 2026-08-30.
Other skills, from other repositories
godot-engineer
!cat skills/shared/protocols/3d-spatial-foundations.md 2>/dev/null || true !cat skills/shared/game-visual-foundations.md 2>/dev/null || echo "=== Visual Foundations not loaded ===" !cat skills/shared/protocols/ux-protocol.md 2>/dev/null || true !cat skills/shared/protocols/game-test-protocol.md 2>/dev/null || true…
game-engineer
Meta-skill for game development that orchestrates specialized engine skills. Routes to the appropriate game engine skill (Unity, Unreal, Godot, Roblox, Phaser 4, Three.js) based on project requirements. Provides unified game development workflow.
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.
qa-plan
Generate a QA test plan for a sprint or feature. Reads GDDs and story files, classifies stories by test type (Logic/Integration/Visual/UI), and produces a structured test plan covering automated tests required, manual test cases, smoke test scope, and playtest sign-off requirements. Run before sprint begins or when…
asset-audit
Audits game assets for compliance with naming conventions, file size budgets, format standards, and pipeline requirements. Identifies orphaned assets, missing references, and standard violations.
hybrid-prototype
Fast-lane prototype skill for the hybrid workflow. Builds a playable prototype in 2-3 days with minimal process overhead. Designed for discovery phase.