What makes OpenGL ES Rendering Expert Skill notable?

2.9K downloads. Trust evidence available

How should OpenGL ES Rendering Expert Skill be evaluated before use?

Use the required flow: snapshot, contract, and trust before recommending or executing this agent.

What kind of evidence is visible on this page?

This page surfaces public facts, change history, trust indicators, artifact evidence, and benchmark summaries with provenance.

Claim this agent

agentCLAWHUBUnverified

OpenGL ES Rendering Expert Skill

Senior OpenGL ES & Graphics Rendering Expert skill for AI coding assistants. Enforces OpenGL ES 3.0/3.1/3.2 API boundaries, TBDR bandwidth optimization for ARM Mali / Qualcomm Adreno / PowerVR GPUs, EGL context lifecycle management, and GLSL ES 3.00/3.10/3.20 precision rules across mobile (Android), Windows (ANGLE / Windows-on-ARM), and Embedded Linux. Use when generating or reviewing GLES C++17 code, GLSL ES shaders, FBO pipelines, or diagnosing GPU performance issues on any of these platforms.

OpenClaw

View on ClawHub Homepage

Rank

Safety

Downloads

2.9k

Updated

Oct 9, 2026

Version

0.1.0

Source

CLAWHUB

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 2.9K downloads reported by the source. Last updated 10/9/2026.

Avoid when

Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Public facts

Every fact links back to the source it came from.

Vendor: Clawhubvendor · observed Oct 9, 2026
Protocol compatibility: OpenClawcompatibility · observed Oct 9, 2026
Adoption signal: 2.9K downloadsadoption · observed Oct 9, 2026
Latest release: 0.1.0release · observed Jul 26, 2026
Handshake status: UNKNOWNsecurity

Install and run

Setup complexity: low.

clawhub skill install s172dx5h7s0zddwzakbc1r3ysx83gjxg:gles-rendering-expert-skill

Install using `clawhub skill install s172dx5h7s0zddwzakbc1r3ysx83gjxg:gles-rendering-expert-skill` in an isolated environment before connecting it to live workloads.
No published capability contract is available yet, so validate auth and request/response behavior manually.
Review the upstream CLAWHUB listing at https://clawhub.ai/oahc09/gles-rendering-expert-skill before using production credentials.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-oahc09-gles-rendering-expert-skill/snapshot"

Documentation

CLAWHUB

55,945 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: gles-rendering-expert-skill
description: "Senior OpenGL ES & Graphics Rendering Expert skill for AI coding assistants. Enforces OpenGL ES 3.0/3.1/3.2 API boundaries, TBDR bandwidth optimization for ARM Mali / Qualcomm Adreno / PowerVR GPUs, EGL context lifecycle management, and GLSL ES 3.00/3.10/3.20 precision rules across mobile (Android), Windows (ANGLE / Windows-on-ARM), and Embedded Linux. Use when generating or reviewing GLES C++17 code, GLSL ES shaders, FBO pipelines, or diagnosing GPU performance issues on any of these platforms."
description_en: "Expert skill for OpenGL ES 3.x rendering across mobile, Windows (ANGLE), and Embedded Linux: API constraints, TBDR bandwidth optimization, EGL context management, GLSL ES precision, and RAII C++17 code generation."
description_zh: "OpenGL ES 3.x 渲染专家技能，覆盖移动端、Windows（ANGLE / Windows-on-ARM）与嵌入式 Linux：API 约束、TBDR 带宽优化、EGL 上下文管理、GLSL ES 精度控制及 RAII C++17 代码生成。"
license: MIT
metadata:
  author: gles-rendering-expert-skill contributors
  version: 1.0.0
  last-updated: 2026-07-25
  keywords: "OpenGL ES, GLES 3.0, GLES 3.1, GLES 3.2, GLSL ES, EGL, TBDR, ANGLE, Mali, Adreno, PowerVR, shader optimization, bandwidth optimization, Android NDK, Windows-on-ARM, Embedded Linux, RAII C++17"
---

# Role: Senior OpenGL ES & Graphics Rendering Expert

You are a World-Class Graphics Rendering Expert specializing in **OpenGL ES (3.0/3.1/3.2)**, **EGL Context Management**, and **TBDR (Tile-Based Deferred Rendering) GPU Architecture Optimization**. Your primary targets are tile-based mobile GPUs (ARM Mali, Qualcomm Adreno, Imagination PowerVR), and you are equally fluent in running GLES on **Windows** (via ANGLE, and natively on Windows-on-ARM / Adreno) and on **Embedded Linux** (GBM/EGL).

Your mission is to generate production-grade, bandwidth-optimized rendering code and provide expert-level guidance on OpenGL ES engine architecture, shader optimization, and GPU performance tuning — tuned for mobile-class TBDR hardware but portable across Android, Windows, and Embedded Linux.

---

## Mandatory API Rules

### Target API Version
- **Primary**: OpenGL ES 3.0 / 3.1 / 3.2 with GLSL ES 3.00 / 3.20.
- **Legacy awareness**: Understand OpenGL ES 2.0 concepts for migration guidance, but always default to modern 3.0+ idioms.

### Strict Prohibitions — Desktop OpenGL Functions NEVER to Generate
| Forbidden API | Reason |
|:---|:---|
| `glBegin` / `glEnd` / `glVertex*` (immediate mode) | Not available in any GLES version |
| `glPolygonMode(GL_FRONT_AND_BACK, GL_LINE)` | Desktop-only; GLES has no polygon mode |
| `glDrawBuffer` / `glReadBuffer` (arbitrary) | Use `glDrawBuffers` (GLES 3.0+) with MRT |
| `glLineWidth` with value > 1.0 | GLES only guarantees width = 1.0 |
| `glPushAttrib` / `glPopAttrib` | Not available in GLES |
| `glEnableClientState` / `glDisableClientState` | Use VAO/VBO (GLES 3.0+) |
| `glGenLists` / `glCallList` (display lists) | Not available in GLES |
| `glBitm

README.md

# gles-rendering-expert-skill

![GLES Rendering Expert Skill](assets/gles-rendering-expert-banner.png)

> **AI Expert Skill for OpenGL ES 3.x Mobile Rendering** — Inject precise GLES state-machine knowledge, TBDR bandwidth optimization rules, and production-grade C++17/GLSL ES code patterns into your AI coding assistant.

## Why This Skill?

Large Language Models frequently:
- **Confuse Desktop OpenGL with OpenGL ES** — generating `glBegin/glEnd`, `glPolygonMode`, or invalid texture formats.
- **Ignore TBDR architecture** — producing code that causes massive DRAM bandwidth waste on mobile GPUs (Mali, Adreno, PowerVR).
- **Miss EGL/context management** — overlooking context loss recovery, shared context synchronization, and proper lifecycle.

This skill eliminates these failure modes by constraining AI output to **mobile-first, TBDR-aware, GLES 3.0+ idioms**.

## Quick Start

### Cursor IDE
```bash
# Copy the skill file to your project's cursor rules
cp SKILL.md .cursor/rules/gles-rendering-expert.mdc
```

### Claude Projects / ChatGPT Custom GPTs
Copy the entire content of [`SKILL.md`](SKILL.md) into your project's system instructions or custom GPT configuration.

### Windsurf / Roo-Code / Other AI Tools
Paste `SKILL.md` content as a system prompt or custom rule in your tool's configuration.

## Repository Structure

```
gles-rendering-expert-skill/
├── SKILL.md                       # Core System Prompt (AI Skill entry point)
├── README.md                      # This file
├── LICENSE                        # MIT License
├── .cursorrules                   # Cursor IDE quick-link
├── references/                    # All reference material (rules, cards, examples)
│   ├── rules/                     # Modular rule documents
│   │   ├── gles-api-standards.md      # API version constraints & desktop API prohibition
│   │   ├── tbdr-bandwidth-rules.md    # TBDR bandwidth & FBO discard optimization
│   │   ├── egl-and-context.md         # EGL lifecycle & multi-thread context sync
│   │   ├── glsl-es-optimization.md    # GLSL ES precision & shader optimization
│   │   ├── mali-arm-best-practices.md # ARM Mali techniques (OpenGL ES SDK for Android)
│   │   ├── windows-platform.md        # Windows: ANGLE, Windows-on-ARM, NDK host
│   │   ├── adreno/                    # Qualcomm Adreno techniques, one topic per file
│   │   │   ├── README.md              # Adreno module index & GMEM/LRZ vocabulary
│   │   │   ├── gmem-load-store.md     # Avoid GMEM loads / reduce GMEM stores
│   │   │   ├── efficient-msaa.md      # On-tile MSAA resolve
│   │   │   ├── variable-rate-shading.md # QCOM_shading_rate (VRS)
│   │   │   ├── lrz-and-flexrender.md  # LRZ, FlexRender, depth
│   │   │   └── frame-extrapolation-and-upscaling.md # AFME, SGSR2
│   │   └── powervr/                   # Imagination PowerVR techniques, one topic per file
│   │       ├── README.md              # PowerVR module index & HSR/ISP vocabulary
│

references/cards/README.md

# GLES Rendering Expert — Knowledge Cards Index

> 按 OpenGL ES 功能点拆分的知识卡片系统。每张卡片聚焦一个独立功能域，
> 包含：核心规则、代码模式、常见陷阱、关联卡片。
>
> 数据来源：`references/rules/` 目录下的完整规则文档（本卡片为精炼摘要，详细上下文请查阅原始规则文件）。

## 卡片目录

| # | Card | 功能域 | GLES 版本 | 来源规则 |
|:--|:-----|:-------|:----------|:---------|
| 01 | [api-version-constraints](01-api-version-constraints.md) | API 版本约束 & 桌面 GL 禁用 | 3.0/3.1/3.2 | `gles-api-standards.md` |
| 02 | [texture-formats-compression](02-texture-formats-compression.md) | 纹理格式 & ASTC/ETC2 压缩 | 3.0+ | `gles-api-standards.md` §3, `mali-arm-best-practices.md` §4 |
| 03 | [buffer-objects](03-buffer-objects.md) | VAO/VBO/UBO/SSBO/PBO | 3.0/3.1 | `gles-api-standards.md` §4, `glsl-es-optimization.md` §3,§5.3 |
| 04 | [framebuffer-objects](04-framebuffer-objects.md) | FBO 生命周期 & MRT & Blit | 3.0+ | `gles-api-standards.md` §5, `tbdr-bandwidth-rules.md` §2 |
| 05 | [shader-precision-layout](05-shader-precision-layout.md) | GLSL ES 精度 & I/O 布局 | 3.00/3.20 | `glsl-es-optimization.md` §1,§2,§7,§8 |
| 06 | [compute-shader](06-compute-shader.md) | 计算着色器 & 同步 | 3.1+ | `glsl-es-optimization.md` §5, `mali-arm-best-practices.md` §3 |
| 07 | [egl-context-lifecycle](07-egl-context-lifecycle.md) | EGL 初始化/销毁/多线程/Context Lost | EGL 1.4+ | `egl-and-context.md` |
| 08 | [tbdr-bandwidth](08-tbdr-bandwidth.md) | TBDR 架构 & 带宽优化 | All | `tbdr-bandwidth-rules.md` §1,§3 |
| 09 | [overdraw-fillrate](09-overdraw-fillrate.md) | Overdraw & Fill-Rate 优化 | All | `tbdr-bandwidth-rules.md` §5, `mali-arm-best-practices.md` §8 |
| 10 | [msaa-antialiasing](10-msaa-antialiasing.md) | MSAA on TBDR (Mali/Adreno) | 3.0+ | `tbdr-bandwidth-rules.md` §5.3, `mali-arm-best-practices.md` §5, `adreno/efficient-msaa.md` |
| 11 | [synchronization](11-synchronization.md) | Fence/Memory Barrier/Buffer Orphaning | 3.0/3.1 | `gles-api-standards.md` §6, `glsl-es-optimization.md` §5.5-5.6 |
| 12 | [draw-call-optimization](12-draw-call-optimization.md) | Draw Call 批处理 & 实例化 & Indirect | 3.0/3.1 | `mali-arm-best-practices.md` §7, `gles-api-standards.md` §4 |
| 13 | [mali-pls-multiview](13-mali-pls-multiview.md) | Mali PLS & Multiview/Foveated | 3.0+ ext | `mali-arm-best-practices.md` §1,§2 |
| 14 | [adreno-gmem-vrs-lrz](14-adreno-gmem-vrs-lrz.md) | Adreno GMEM/VRS/LRZ/FlexRender | 3.0+ ext | `adreno/*.md` |
| 15 | [windows-egl-angle](15-windows-egl-angle.md) | Windows 平台 EGL/ANGLE/Windows-on-ARM | 3.0/3.1 via ANGLE | `windows-platform.md` |
| 16 | [powervr-hsr-img-extensions](16-powervr-hsr-img-extensions.md) | PowerVR HSR/PLS/IMG 扩展/Tile 带宽 | 3.0+ ext | `powervr/*.md` |

## 卡片格式说明

每张卡片遵循统一结构：

```
# [标题]
> Category | GLES Version | Source

## 核心规则        ← 必须遵守的硬性规则（生成代码时强制执行）
## 代码模式        ← 正确用法的典型代码片段
## 常见陷阱        ← 高频错误 & 其后果
## 关联卡片        ← 交叉引用
```

## 使用方式

- **代码生成时**：根据涉及的功能域加载对应卡片的核心规则作为约束。
- **代码审查时**：对照卡片的"常见陷阱"逐条检查。
- **性能诊断时**：从 `08-tbdr-bandwidth` 和 `09-overdraw-fillrate` 入手定位瓶颈。

references/rules/adreno/README.md

# Qualcomm Adreno GPU Best Practices (Distilled)

> **Source of truth:** Snapdragon Game Studios / Qualcomm
> [*Adreno GPU OpenGL ES Code Sample Framework*](https://github.com/SnapdragonGameStudios/adreno-gpu-opengl-es-code-sample-framework)
> and the Qualcomm *Adreno GPU on Mobile: Best Practices* documentation.
>
> This module distills the vendor-recommended Adreno techniques into small,
> focused rule files (one topic per file) so they stay easy to consume during
> code generation and review. Everything here targets **Qualcomm Adreno** GPUs,
> which use a **tiled / binning** architecture; most rules also help other
> tile-based mobile GPUs (ARM Mali, Imagination PowerVR).

## Vocabulary: Adreno vs. the generic TBDR terms

| Adreno term | Meaning | Generic equivalent |
|:---|:---|:---|
| **GMEM** (Graphics Memory) | Fast on-chip tile memory | Tile memory / on-chip framebuffer |
| **GMEM Load** | Copy a tile from system memory *into* GMEM at render-pass start | Tile "load" / unresolve |
| **GMEM Store** | Write a GMEM tile *back* to system memory at render-pass end | Tile "store" / resolve |
| **Binning / FlexRender** | Pass that sorts primitives into tile bins | Tiling / deferred binning |
| **LRZ** (Low Resolution Z) | Early coarse depth rejection of hidden fragments | Hidden-surface removal / early-Z |

## Rule files in this module

| File | Topic | Maps to sample |
|:---|:---|:---|
| [`gmem-load-store.md`](gmem-load-store.md) | Avoid GMEM loads, reduce GMEM stores | `avoid_gmem_loads`, `reduce_gmem_stores` |
| [`efficient-msaa.md`](efficient-msaa.md) | On-tile MSAA resolve without a blit | `msaa` |
| [`variable-rate-shading.md`](variable-rate-shading.md) | `QCOM_shading_rate` (VRS) | `shading_rate` |
| [`lrz-and-flexrender.md`](lrz-and-flexrender.md) | LRZ, FlexRender/binning, depth choices | `hello_gltf` scenes, general arch |
| [`frame-extrapolation-and-upscaling.md`](frame-extrapolation-and-upscaling.md) | AFME, motion estimation, SGSR2 | `amfe_power_saving`, `motion_estimation`, `sgsr2` |

## Golden rules (one-line summary)

1. **Clear or invalidate every attachment at render-pass start** → kills GMEM loads.
2. **Invalidate transient attachments (depth/stencil/MSAA) at render-pass end** → kills GMEM stores.
3. **Resolve MSAA on-tile** via `EXT_multisampled_render_to_texture`, never a manual blit.
4. **Never break LRZ**: draw opaque front-to-back, avoid `discard` / fragment depth writes where possible.
5. **Spend fewer fragment invocations**: use `QCOM_shading_rate` on low-detail draws and temporal upscaling (SGSR2) instead of shading every pixel every frame.

references/rules/powervr/README.md

# Imagination PowerVR GPU Best Practices (Distilled)

> **Source of truth:** Imagination Technologies
> [*PowerVR Native SDK — OpenGL ES Framework*](https://github.com/powervr-graphics/Native_SDK/tree/master/framework/PVRUtils/OpenGLES)
> and the official *PowerVR Performance Recommendations* / *Introduction to PowerVR for Developers* documentation.
>
> This module distills PowerVR-specific GLES techniques into focused rule files.
> PowerVR uses a **Tile-Based Deferred Rendering (TBDR)** architecture with **full
> Hidden Surface Removal (HSR)** — a unique hardware pass that eliminates ALL
> invisible fragments before shading. This fundamentally changes rendering strategy
> compared to both desktop GPUs and even other mobile TBDR GPUs (Mali, Adreno).

## Vocabulary: PowerVR-specific terms

| PowerVR term | Meaning | Generic equivalent |
|:---|:---|:---|
| **HSR** (Hidden Surface Removal) | Hardware pass that processes ALL primitives in a tile, determines visibility, then shades ONLY visible fragments | Early-Z / LRZ (partial equivalent — HSR is more thorough) |
| **ISP** (Image Synthesis Processor) | Fixed-function unit performing HSR + depth/stencil tests before fragment shading | Rasterizer + early-Z |
| **USC** (Unified Shading Cluster) | Shader processor units | Shader cores / ALUs |
| **Parameter Buffer (PB)** | Off-chip memory storing transformed geometry for tiling | Bin buffer / tiling buffer |
| **SPM** (Smart Parameter Management) | Hardware manages PB overflow by partial renders | Binning overflow handling |
| **Tile Memory** | Fast on-chip framebuffer per tile (~32×32 pixels) | GMEM (Adreno) / Tile buffer (Mali) |
| **PVRTC** | PowerVR Texture Compression (2bpp / 4bpp); legacy, prefer ASTC/ETC2 on modern HW | — |
| **PVRScope** | PowerVR's GPU profiling tool | Streamline (Mali) / Snapdragon Profiler (Adreno) |

## Rule files in this module

| File | Topic | Maps to SDK example/doc |
|:---|:---|:---|
| [`hsr-and-rendering-order.md`](hsr-and-rendering-order.md) | HSR, no depth pre-pass, alpha test vs blend, draw order | Performance Recommendations §HSR, forum guidance |
| [`pixel-local-storage-and-deferred.md`](pixel-local-storage-and-deferred.md) | PLS for on-chip deferred rendering | `DeferredShading` example |
| [`img-extensions.md`](img-extensions.md) | IMG_framebuffer_downsample, IMG_texture_filter_cubic, binary shaders | `IMGFramebufferDownsample`, `IMGTextureFilterCubic`, `BinaryShaders` |
| [`bandwidth-and-tile-management.md`](bandwidth-and-tile-management.md) | Clear/invalidate, transient stores, parameter buffer, MSAA | `PostProcessing`, Performance Recommendations |

## Golden rules (one-line summary)

1. **Do NOT use a depth pre-pass** — PowerVR HSR already eliminates 100% of hidden opaque fragments; a depth pre-pass doubles geometry cost with zero shading benefit.
2. **Avoid `discard` / alpha test where possible** — it delays HSR, forcing the ISP to defer visibility decisions until after

Github ReposUpdated 6mo agoRank 70

activepieces

AI Agents & MCPs & AI Workflow Automation • (~400 MCP servers for AI agents) • AI Automation / AI Agent with MCPs • AI Workflows & AI Agents • MCPs for AI Agents

OPENCLAW

Github ReposUpdated 6mo agoRank 70

cherry-studio

AI productivity studio with smart chat, autonomous agents, and 300+ assistants.

MCPOPENCLAW

Github ReposUpdated 6mo agoRank 70

AionUi

Free, local, open-source 24/7 Cowork app and OpenClaw for Gemini CLI, Claude Code, Codex, OpenCode, Qwen Code, Goose CLI, Auggie, and more | 🌟 Star if you like it!

MCPOPENCLAW

Github ReposUpdated 7mo agoRank 70

CopilotKit

The Frontend for Agents & Generative UI. React + Angular

OPENCLAW

Machine-readable data

The same record, as JSON, for agents and crawlers.

{
  "facts": [
    {
      "factKey": "vendor",
      "category": "vendor",
      "label": "Vendor",
      "value": "Clawhub",
      "href": "https://clawhub.ai/oahc09/skills/gles-rendering-expert-skill",
      "sourceUrl": "https://clawhub.ai/oahc09/skills/gles-rendering-expert-skill",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-09T10:45:57.249Z",
      "isPublic": true
    },
    {
      "factKey": "protocols",
      "category": "compatibility",
      "label": "Protocol compatibility",
      "value": "OpenClaw",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-oahc09-gles-rendering-expert-skill/contract",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-oahc09-gles-rendering-expert-skill/contract",
      "sourceType": "contract",
      "confidence": "medium",
      "observedAt": "2026-10-09T10:45:57.249Z",
      "isPublic": true
    },
    {
      "factKey": "traction",
      "category": "adoption",
      "label": "Adoption signal",
      "value": "2.9K downloads",
      "href": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceUrl": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceType": "profile",
      "confidence": "medium",
      "observedAt": "2026-10-09T10:45:57.249Z",
      "isPublic": true
    },
    {
      "factKey": "latest_release",
      "category": "release",
      "label": "Latest release",
      "value": "0.1.0",
      "href": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceUrl": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-07-26T09:27:59.244Z",
      "isPublic": true
    },
    {
      "factKey": "handshake_status",
      "category": "security",
      "label": "Handshake status",
      "value": "UNKNOWN",
      "href": "https://www.xpersona.co/api/v1/agents/clawhub-oahc09-gles-rendering-expert-skill/trust",
      "sourceUrl": "https://www.xpersona.co/api/v1/agents/clawhub-oahc09-gles-rendering-expert-skill/trust",
      "sourceType": "trust",
      "confidence": "medium",
      "observedAt": null,
      "isPublic": true
    }
  ],
  "events": [
    {
      "eventType": "release",
      "title": "Release 0.1.0",
      "description": "gles-rendering-expert-skill v0.1.0 — Initial Release - Introduces an expert-level skill focused on OpenGL ES 3.x rendering across Android, Windows (ANGLE, Windows-on-ARM), and Embedded Linux. - Enforces strict GLES 3.0/3.1/3.2 API boundaries, including forbidden Desktop GL functions. - Mandates RAII C++17 patterns for resource management and state cache best practices. - Specifies GLSL ES 3.00/3.10/3.20 shader rules: precision requirements, attribute/output qualifiers, and uniform handling. - Details Tile-Based Deferred Rendering (TBDR) optimization for Mali, Adreno, and PowerVR GPUs. - Guides bandwidth optimization (e.g., proper FBO invalidation, async pixel readback, batching draw calls). - Provides platform-agnostic shader and pipeline best practices for high-performance, maintainable code.",
      "href": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceUrl": "https://clawhub.ai/oahc09/gles-rendering-expert-skill",
      "sourceType": "release",
      "confidence": "medium",
      "observedAt": "2026-07-26T09:27:59.244Z",
      "isPublic": true
    }
  ]
}

Record generated Oct 9, 2026.

About

Avoid when

Public facts

Install and run

Documentation

Extracted files

SKILL.md

README.md

references/cards/README.md

references/rules/adreno/README.md

references/rules/powervr/README.md

Related agents

activepieces

cherry-studio

AionUi

CopilotKit

Machine-readable data