Borrowing it
Nothing to install: this file belongs to tanepiper/unbanished. 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/tanepiper/unbanished/main/.agents/skills/banished-shader/SKILL.mdgit clone --depth 1 https://github.com/tanepiper/unbanishedWrote 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/tanepiper/unbanished/banished-shader)<a href="https://agentmods.dev/skills/tanepiper/unbanished/banished-shader"><img src="https://agentmods.dev/badge/skills/tanepiper/unbanished/banished-shader/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/tanepiper/unbanished/banished-shader"><img src="https://agentmods.dev/badge/skills/tanepiper/unbanished/banished-shader.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 | $0.00074 | $0.00730 |
| Opus 5.5 | $0.00030 | $0.00292 |
| Sonnet 5.5 | $0.00015 | $0.00146 |
| Haiku 4.5 | $0.00007 | $0.00073 |
Grade A, and why
banished-shader 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 3d 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 — 49 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Port a shader to WGSL
1. Find and disassemble
Materials reference programs by resource path (decode the MaterialInstance/material with
banished-format tools to find them). Disassemble the shipped D3D9 bytecode:
python3 tools/shader/d3d9_disasm.py game-data/extracted/<path>/X_vs.crs # list variants
python3 tools/shader/d3d9_disasm.py game-data/extracted/<path>/X_ps.crs 0 > analysis/X_ps0.txt
Each resource holds several D3D9 variants then DXBC blobs; the CTAB lists constant names and
registers — that tells you what each c# is (view_proj, light_dir, fog…). Pick the variant
the game uses (usually 0; confirm via the material's program index). Keep dumps in analysis/.
2. Translate
- Translate instruction by instruction first, then simplify only where provably identical.
Keep the original constants verbatim (e.g.
0.00390625,-0.666666687). - D3D9 → wgpu differences:
- clip-space depth is 0..1 in both; but matrices are row-vector in D3D9 (
m4x4= dot rows) — check transpose against the CPU side; texkill→discard;sincos,lrp(a,b,c)=mix(c, b, a);sge/slt→select/step;- saturate modifiers
_sat→clamp(x, 0.0, 1.0).
- clip-space depth is 0..1 in both; but matrices are row-vector in D3D9 (
- World space is the terrain's: x/y in cells (map-centred), z up = saved height / 256.
- Prepend the shared
GLOBALS_WGSL(scene_gfx.rs) instead of redeclaringGlobals; callreflect_clip(world_z)for the planar reflection pass.
3. WGSL / WebGL2 rules (learned the hard way)
textureSampleneeds uniform control flow: do every sample before anydiscard(putreflect_clip()/ alpha-test discards at the end).- WebGL2 has no base-vertex draws, limited storage buffers in vertex stages, and needs
downlevel_webgl2_defaultslimits — use vertex/index buffers and uniforms. - Render targets use
SCENE_FORMAT(Rgba8Unorm) andDEPTH_FORMATfromscene_gfx.rs. - To preserve destination alpha (e.g. decals over terrain) use an alpha blend of
Zero/One.
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.
- 3d ago First seen · 49 lines · 74 tokens per session scan A ed860dc4cb90
banished-shader is a skill published in the GitHub repository tanepiper/unbanished (23 stars, last pushed 22d ago), licensed Apache-2.0. It adds 74 tokens to every session and 730 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 2026-11-15.
banished-shader compared with skills from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| banished-shadertanepiper/unbanished | Skill | 23 | 22d ago | Apache-2.0 | A | 74 | Claude Code, Codex |
| soundclawiamlukethedev/Claw3D | Skill | 2.5k | 14d ago | MIT | A | 18 | Claude Code, Codex |
| task-manageriamlukethedev/Claw3D | Skill | 2.5k | 14d ago | MIT | A | 56 | Claude Code, Codex |
| todoiamlukethedev/Claw3D | Skill | 2.5k | 14d ago | MIT | A | 12 | Claude Code, Codex |
| gamedev-shippingericrisco/rsc-harness | Skill | 636 | today | MIT | A | 99 | Claude Code, Codex |
| self-evolvetettethu/VibeGame | Skill | 388 | 10d ago | Apache-2.0 | A | 50 | Claude Code, Codex |
| artist-self-evolvetettethu/VibeGame | Skill | 388 | 10d ago | Apache-2.0 | A | 62 | Claude Code, Codex |