slug-font-rendering
Reference HLSL shader implementations for the Slug font rendering algorithm, enabling high-quality GPU-accelerated vector font and glyph rendering.
What this skill does
# Slug Font Rendering Algorithm
> Skill by [ara.so](https://ara.so) — Daily 2026 Skills collection.
Slug is a reference implementation of the Slug font rendering algorithm — a GPU-accelerated technique for rendering vector fonts and glyphs at arbitrary scales with high quality anti-aliasing. It works by encoding glyph outlines as lists of quadratic Bézier curves and line segments, then resolving coverage directly in fragment shaders without pre-rasterized textures.
**Paper:** [JCGT 2017 — Slug Algorithm](https://jcgt.org/published/0006/02/02/)
**Blog (updates):** [A Decade of Slug](https://terathon.com/blog/decade-slug.html)
**License:** MIT — Patent dedicated to public domain. Credit required if distributed.
---
## What Slug Does
- Renders TrueType/OpenType glyphs entirely on the GPU
- No texture atlases or pre-rasterization needed
- Scales to any resolution without quality loss
- Anti-aliased coverage computed per-fragment using Bézier math
- Works with any rendering API that supports programmable shaders (D3D11/12, Vulkan, Metal via translation)
---
## Repository Structure
```
Slug/
├── slug.hlsl # Core fragment shader — coverage computation
├── band.hlsl # Band-based optimization for glyph rendering
├── curve.hlsl # Quadratic Bézier and line segment evaluation
├── README.md
```
---
## Installation / Integration
Slug is a **reference implementation** — you integrate the HLSL shaders into your own rendering pipeline.
### Step 1: Clone the Repository
```bash
git clone https://github.com/EricLengyel/Slug.git
```
### Step 2: Include the Shaders
Copy the `.hlsl` files into your shader directory and include them in your pipeline:
```hlsl
#include "slug.hlsl"
#include "curve.hlsl"
```
### Step 3: Prepare Glyph Data on the CPU
You must preprocess font outlines (TrueType/OTF) into Slug's curve buffer format:
- Decompose glyph contours into quadratic Bézier segments and line segments
- Upload curve data to a GPU buffer (structured buffer or texture buffer)
- Precompute per-glyph "band" metadata for the band optimization
---
## Core Concepts
### Glyph Coordinate System
- Glyph outlines live in **font units** (typically 0–2048 or 0–1000 per em)
- The fragment shader receives a position in glyph space via interpolated vertex attributes
- Coverage is computed by counting signed curve crossings in the Y direction (winding number)
### Curve Data Format
Each curve entry in the GPU buffer stores:
```hlsl
// Line segment: p0, p1
// Quadratic Bézier: p0, p1 (control), p2
struct CurveRecord
{
float2 p0; // Start point
float2 p1; // Control point (or end point for lines)
float2 p2; // End point (unused for lines — flagged via type)
// Type/flags encoded separately or in padding
};
```
### Band Optimization
The glyph bounding box is divided into horizontal **bands**. Each band stores only the curves that intersect it, reducing per-fragment work from O(all curves) to O(local curves).
---
## Key Shader Code & Patterns
### Fragment Shader Entry Point (Conceptual Integration)
```hlsl
// Inputs from vertex shader
struct PS_Input
{
float4 position : SV_Position;
float2 glyphCoord : TEXCOORD0; // Position in glyph/font units
// Band index or precomputed band data
nointerpolation uint bandOffset : TEXCOORD1;
nointerpolation uint curveCount : TEXCOORD2;
};
// Glyph curve data buffer
StructuredBuffer<float4> CurveBuffer : register(t0);
float4 PS_Slug(PS_Input input) : SV_Target
{
float coverage = ComputeGlyphCoverage(
input.glyphCoord,
CurveBuffer,
input.bandOffset,
input.curveCount
);
// Premultiplied alpha output
float4 color = float4(textColor.rgb * coverage, coverage);
return color;
}
```
### Quadratic Bézier Coverage Computation
The heart of the algorithm — computing signed coverage from a quadratic Bézier:
```hlsl
// Evaluate whether a quadratic bezier contributes to coverage at point p
// p0: start, p1: control, p2: end
// Returns signed coverage contribution
float QuadraticBezierCoverage(float2 p, float2 p0, float2 p1, float2 p2)
{
// Transform to canonical space
float2 a = p1 - p0;
float2 b = p0 - 2.0 * p1 + p2;
// Find t values where bezier Y == p.y
float2 delta = p - p0;
float A = b.y;
float B = a.y;
float C = p0.y - p.y;
float coverage = 0.0;
if (abs(A) > 1e-6)
{
float disc = B * B - A * C;
if (disc >= 0.0)
{
float sqrtDisc = sqrt(disc);
float t0 = (-B - sqrtDisc) / A;
float t1 = (-B + sqrtDisc) / A;
// For each valid t in [0,1], compute x and check winding
if (t0 >= 0.0 && t0 <= 1.0)
{
float x = (A * t0 + 2.0 * B) * t0 + p0.x + delta.x;
// ... accumulate signed coverage
}
if (t1 >= 0.0 && t1 <= 1.0)
{
float x = (A * t1 + 2.0 * B) * t1 + p0.x + delta.x;
// ... accumulate signed coverage
}
}
}
else
{
// Degenerate to linear case
float t = -C / (2.0 * B);
if (t >= 0.0 && t <= 1.0)
{
float x = 2.0 * a.x * t + p0.x;
// ... accumulate signed coverage
}
}
return coverage;
}
```
### Line Segment Coverage
```hlsl
// Signed coverage contribution of a line segment from p0 to p1
float LineCoverage(float2 p, float2 p0, float2 p1)
{
// Check Y range
float minY = min(p0.y, p1.y);
float maxY = max(p0.y, p1.y);
if (p.y < minY || p.y >= maxY)
return 0.0;
// Interpolate X at p.y
float t = (p.y - p0.y) / (p1.y - p0.y);
float x = lerp(p0.x, p1.x, t);
// Winding: +1 if p is to the left (inside), -1 if right
float dir = (p1.y > p0.y) ? 1.0 : -1.0;
return (p.x <= x) ? dir : 0.0;
}
```
### Anti-Aliasing with Partial Coverage
For smooth edges, use the distance to the nearest curve for sub-pixel anti-aliasing:
```hlsl
// Compute AA coverage using partial pixel coverage
// windingNumber: integer winding from coverage pass
// distToEdge: signed distance to nearest curve (in pixels)
float AntiAliasedCoverage(int windingNumber, float distToEdge)
{
// Non-zero winding rule
bool inside = (windingNumber != 0);
// Smooth transition at edges using clamp
float edgeCoverage = clamp(distToEdge + 0.5, 0.0, 1.0);
return inside ? edgeCoverage : (1.0 - edgeCoverage);
}
```
---
## Vertex Shader Pattern
```hlsl
struct VS_Input
{
float2 position : POSITION; // Glyph quad corner in screen/world space
float2 glyphCoord : TEXCOORD0; // Corresponding glyph-space coordinate
uint bandOffset : TEXCOORD1; // Offset into curve buffer for this glyph
uint curveCount : TEXCOORD2; // Number of curves in band
};
struct VS_Output
{
float4 position : SV_Position;
float2 glyphCoord : TEXCOORD0;
nointerpolation uint bandOffset : TEXCOORD1;
nointerpolation uint curveCount : TEXCOORD2;
};
VS_Output VS_Slug(VS_Input input)
{
VS_Output output;
output.position = mul(float4(input.position, 0.0, 1.0), WorldViewProjection);
output.glyphCoord = input.glyphCoord;
output.bandOffset = input.bandOffset;
output.curveCount = input.curveCount;
return output;
}
```
---
## CPU-Side Data Preparation (Pseudocode)
```cpp
// 1. Load font file and extract glyph outlines
FontOutline outline = LoadGlyphOutline(font, glyphIndex);
// 2. Decompose to quadratic Beziers (TrueType is already quadratic)
// OTF cubic curves must be approximated/split into quadratics
std::vector<SlugCurve> curves = DecomposeToQuadratics(outline);
// 3. Compute bands
float bandHeight = outline.bounds.height / NUM_BANDS;
std::vector<BandData> bands = ComputeBands(curves, NUM_BANDS, bandHeight);
// 4. Upload to GPU
UploadStructuredBuffer(curveBuffer, curves.data(), curves.size());
UploadStructurRelated in AI Agents
skill-development
IncludedComprehensive meta-skill for creating, managing, validating, auditing, and distributing Claude Code skills and slash commands (unified in v2.1.3+). Provides skill templates, creation workflows, validation patterns, audit checklists, naming conventions, YAML frontmatter guidance, progressive disclosure examples, and best practices lookup. Use when creating new skills, validating existing skills, auditing skill quality, understanding skill architecture, needing skill templates, learning about YAML frontmatter requirements, progressive disclosure patterns, tool restrictions (allowed-tools), skill composition, skill naming conventions, troubleshooting skill activation issues, creating custom slash commands, configuring command frontmatter, using command arguments ($ARGUMENTS, $1, $2), bash execution in commands, file references in commands, command namespacing, plugin commands, MCP slash commands, Skill tool configuration, or deciding between skills vs slash commands. Delegates to docs-management skill for official documentation.
reprompter
IncludedTransform messy prompts into well-structured, effective prompts — single or multi-agent. Use when: "reprompt", "reprompt this", "clean up this prompt", "structure my prompt", rough text needing XML tags and best practices, "reprompter teams", "repromptception", "run with quality", "smart run", "smart agents", multi-agent tasks, audits, parallel work, anything going to agent teams. Don't use when: simple Q&A, pure chat, immediate execution-only tasks. See "Don't Use When" section for details. Outputs: Structured XML/Markdown prompt, quality score (before/after), optional team brief + per-agent sub-prompts, agent team output files. Success criteria: Single mode quality score ≥ 7/10; Repromptception per-agent prompt quality score 8+/10; all required sections present, actionable and specific.
adaptive-compaction
IncludedAdaptive add-on policy and recovery layer that decides WHEN to compact, prune, snapshot, or fork -- replacing fixed-percent auto-compaction across Claude Code, Codex, and MCP-capable hosts. Trigger on auto-compact timing or damage: "when should I compact", "is it safe to compact now or start a fresh session", "auto-compact fires too early/mid-task", "switching to an unrelated task but the window still has space", "context rot", "answers get worse the longer the session runs", "the agent forgot the plan or my decisions after it summarized", "add a layer on top that manages context without changing the agent", raising autoCompactWindow to give the policy room, or installing/tuning a cross-tool compaction policy or PreCompact hook -- even when "compaction" is never said but the problem is context-window pressure or post-summarization memory loss. Do NOT use to summarize a conversation, build RAG, write a summarization prompt (decides WHEN not HOW), or answer max-context-length trivia.
agent-skill-creator
IncludedCreate cross-platform agent skills from workflow descriptions. Activates when users ask to create an agent, automate a repetitive workflow, create a custom skill, or need advanced agent creation. Triggers on phrases like create agent for, automate workflow, create skill for, every day I have to, daily I need to, turn process into agent, need to automate, create a cross-platform skill, validate this skill, export this skill, migrate this skill. Supports single skills, multi-agent suites, transcript processing, template-based creation, interactive configuration, cross-platform export, and spec validation.
llm-wiki
IncludedUse when building or maintaining a persistent personal knowledge base (second brain) in Obsidian where an LLM incrementally ingests sources, updates entity/concept pages, maintains cross-references, and keeps a synthesis current. Triggers include "second brain", "Obsidian wiki", "personal knowledge management", "ingest this paper/article/book", "build a research wiki", "compound knowledge", "Memex", or whenever the user wants knowledge to accumulate across sessions instead of being re-derived by RAG on every query.
skill-master
IncludedAgent Skills authoring, evaluation, and optimization. Create, edit, validate, benchmark, and improve skills following the agentskills.io specification. Use when designing SKILL.md files, structuring skill folders (references, scripts, assets), ingesting external documentation into skills, running trigger evals, benchmarking skill quality, optimizing descriptions, or performing blind A/B comparisons. Keywords: agentskills.io, SKILL.md, skill authoring, eval, benchmark, trigger optimization.