agentleFS
jaccenjaccen/Awesome-Gaussian-Skills/skills/3dgs-engineering-guide/SKILL.md

3dgs-engineering-guide

Skill186 starsChanged 57 days agoApache-2.0View on GitHub

Guide for deploying 3DGS from research to production: 10 industry verticals, engineering stack, GIS

The file

We split this skill into the rules it gives an agent

10 rules. About 0 are common, 9 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈9

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the user does not specify, defaults are: domain=all, focus=all.

    line 63
  2. Distinctive:

    Read these files from static/:

    line 68
  3. Very specific:

    All domain values load static/industry-landscape.md (the landscape fragment covers all 10 verticals + world models).

    line 72
  4. Distinctive:

    After loading the required fragments:

    line 92
  5. Distinctive:

    No invented data: Never fabricate deployment metrics, VRAM numbers, or system requirements not in the loaded reference files.

    line 114
  6. Distinctive:

    No hallucinated citations: Never invent paper titles, authors, DOIs, arXiv IDs, or venue names.

    line 115
  7. Distinctive:

    No silent speculation: If you are uncertain about a technical detail, explicitly flag it with "[UNCERTAIN]" rather than presenting it as …

    line 116
  8. Distinctive:

    No method misattribution: Do not assign features, results, or mechanisms from one method to another.

    line 117
  9. Distinctive:

    No oversimplified recommendations: Do not reduce complex deployment trade-offs to a single "use this" judgment without considering the …

    line 118
  10. Distinctive:

    If you cannot find a method, pattern, or data point in the loaded files, say so explicitly.

    line 131

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

3DGS Engineering Guide (Router)

Architecture: Axis-driven static/dynamic router. Do NOT try to apply the engineering guidance from memory or from this router alone. Always load fragments from disk as described below.

Step 1 — Detect Request Axes

Analyze the user's request to determine axis values:

Axis: domain

User Intentdomain value
Autonomous driving simulation, sensor simautonomous-driving
Digital twin, smart city, GISdigital-twin
Cultural heritage, museum, 3D archivingculture
Film, game production, virtual productionfilm-games
E-commerce, 3D product display, WebARecommerce
Industrial inspection, drone surveyindustrial
AR/VR/MR headsets, spatial computingar-vr
BIM, architecture, as-built verificationbim
Robotics, embodied AI, manipulationrobotics
Military simulation, defensemilitary
World model, interactive 3D worldworld-model
Unspecified or broad engineering questionall

Axis: focus

User Intentfocus value
Data acquisition, reconstruction, deployment pipelinepipeline
Quality assurance, scalability, cross-platformbest-practices
Debugging artifacts, OOM, deployment failurespitfalls
Choosing tools by use case/platform/scaledecision-tree
Specific technology stack info (formats, compression, engines)tech-stack
GIS integration, spatial analysis, S3M/Cesiumgis-toolchain
Unspecified or comprehensive engineering guidanceall

If the user does not specify, defaults are: domain=all, focus=all.

Step 2 — Load Required Fragments

Always Load (every invocation)

Read these files from static/:

  • static/core-stance.md — Role, workflow, deployment scale reference, terminology, guardrails

On-Demand Load (by detected domain)

All domain values load static/industry-landscape.md (the landscape fragment covers all 10 verticals + world models). The Router will focus attention on the relevant section within the loaded fragment.

On-Demand Load (by detected focus)

focusFragment(s) to Load
pipelinestatic/tech-stack.md
best-practicesstatic/best-practices.md
pitfallsstatic/pitfalls.md
decision-treestatic/decision-trees.md
tech-stackstatic/tech-stack.md
gis-toolchainstatic/gis-toolchain.md
allAll focus fragments (tech-stack, best-practices, pitfalls, decision-trees, gis-toolchain)

Reference Load (when citing specific methods)

  • static/reference-papers.md — Method tables organized by domain

Optimization: For a focused question (e.g., "how to deploy 3DGS on Quest 3"), load core-stance + industry-landscape (domain=ar-vr section) + tech-stack. For broad planning (e.g., "start a digital twin project"), load all fragments.

Step 3 — Execute Engineering Guidance

After loading the required fragments:

  1. Identify use case from domain axis — focus on the relevant industry section
  2. Recommend pipeline from tech-stack fragment
  3. Reference specific methods from reference-papers fragment and knowledge base
  4. Provide concrete next steps — actionable items, not generic advice
  5. Warn about pitfalls from pitfalls fragment — highlight domain-specific failure modes

Cross-Skill Routing

This skill focuses on production engineering. For related workflows:

  • Method comparison → 3dgs-method-compare (for selecting methods for deployment)
  • Code review/bug detection → 3dgs-code-reviewer (for detecting deployment-critical bugs)
  • MCP rendering → 3dgs-mcp-renderer (for real-time rendering integration)
  • Spatial intelligence → 3dgs-spatial-agent (for agent-driven deployment scenarios)
  • CAD/Mesh integration → cad-mesh-3dgs (for BIM/CAD workflows)

Red Lines

The following are categorical prohibitions. Violating any of these invalidates the output:

  • No invented data: Never fabricate deployment metrics, VRAM numbers, or system requirements not in the loaded reference files. If a value is not found, write "data not available" or "N/A".
  • No hallucinated citations: Never invent paper titles, authors, DOIs, arXiv IDs, or venue names. Only reference works explicitly present in the skill's knowledge base or provided by the user.
  • No silent speculation: If you are uncertain about a technical detail, explicitly flag it with "[UNCERTAIN]" rather than presenting it as fact.
  • No method misattribution: Do not assign features, results, or mechanisms from one method to another. Each method's data is specific to that method.
  • No oversimplified recommendations: Do not reduce complex deployment trade-offs to a single "use this" judgment without considering the target environment.

Related Skills

  • 3dgs-method-compare — Method comparison (use for selecting methods for deployment)
  • 3dgs-code-reviewer — Code review (use for detecting deployment-critical bugs)
  • 3dgs-mcp-renderer — MCP rendering protocol (use for real-time rendering integration)
  • 3dgs-spatial-agent — Spatial intelligence (use for agent-driven deployment scenarios)

Guardrail: Do Not Apply From Memory

Do NOT try to apply the logic, method data, bug patterns, or technical details described in this skill from memory. Always read the SKILL.md and referenced files from disk before producing any output. The knowledge base is updated frequently; stale memory may produce outdated, inaccurate, or fabricated results.

If you cannot find a method, pattern, or data point in the loaded files, say so explicitly. Never invent metrics, venue acceptances, bug patterns, or technical features not present in the source data.

If you like it, please star this repo https://github.com/jaccen/Awesome-Gaussian-Skills

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.