3dgs-method-compare
Skill186 starsChanged 57 days agoApache-2.0View on GitHub
Compare 3D Gaussian Splatting variants across 11 dimensions. Built-in knowledge of 1066 methods across
In this repository19 other files it gives its agents.
- awesome-gaussian-skillsCLAUDE.md
- awesome-gaussian-skillsSkill
- 3dgs-articulated-reasonerSkill
- 3dgs-code-reviewerSkill
- 3dgs-compression-deploySkill
- 3dgs-engineering-guideSkill
- 3dgs-experiment-plannerSkill
- 3dgs-mcp-rendererSkill
- 3dgs-paper-readerSkill
- 3dgs-spatial-agentSkill
- 3dgs-training-debuggerSkill
- 3dgs-visualizerSkill
- cad-mesh-3dgsSkill
- cg-paper-writingSkill
- nerf-to-3dgs-migratorSkill
- ontologySkill
- patent-software-ipSkill
- research-expert-teamSkill
- super-agentSkill
The file
We split this skill into the rules it gives an agent
12 rules. About 0 are common, 12 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈12
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 52
If the user does not specify, defaults are: category=all, depth=standard.
- Distinctive: line 57
Read these files from static/:
- Distinctive: line 61
Read the corresponding fragment(s) from static/:
- Distinctive: line 80
After loading the required fragments:
- Distinctive: line 83
Use method data from the loaded category fragment(s)
- Distinctive: line 84
Follow output format from output-rules.md if producing tables/reports
- Distinctive: line 103
No invented data: Never fabricate benchmark results, timing numbers, or method characteristics not in the loaded reference files.
- Distinctive: line 104
No hallucinated citations: Never invent paper titles, authors, DOIs, arXiv IDs, or venue names.
- Distinctive: line 105
No silent speculation: If you are uncertain about a technical detail, explicitly flag it with "[UNCERTAIN]" rather than presenting it as …
- Distinctive: line 106
No method misattribution: Do not assign features, results, or mechanisms from one method to another.
- Distinctive: line 107
No oversimplified comparisons: Do not reduce multi-dimensional trade-offs to a single "better/worse" judgment without context.
- Distinctive: line 123
If you cannot find a method, pattern, or data point in the loaded files, say so explicitly.
≈ “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.
Files
3DGS Method Comparison Engine (Router)
Architecture: Axis-driven static/dynamic router. Do NOT try to apply the comparison logic 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: category
| User Intent | category value |
| Foundation/baseline methods, opacity, compression | core |
| Surface reconstruction, geometry, text-to-3D generation | geometry |
| Language features, semantic, feed-forward inference | semantic |
| SLAM, large-scale scene, urban reconstruction | slam |
| Dynamic scenes, 4DGS, human/avatar, articulated objects | dynamic |
| Cross-domain, autonomous driving, spatial intelligence, editing, systems | application |
| Broad/unspecified comparison across many categories | all |
Axis: depth
| User Intent | depth value |
| Quick lookup, single method info | quick |
| Standard comparison (2-5 methods, focused analysis) | standard |
| Comprehensive survey, full-category scan | comprehensive |
If the user does not specify, defaults are: category=all, depth=standard.
Step 2 — Load Required Fragments
Always Load (every invocation)
Read these files from static/:
- static/core-stance.md — Role, capabilities, comparison dimensions, rendering comparison table, guardrails
On-Demand Load (by detected category)
Read the corresponding fragment(s) from static/:
| category | Fragment to Load |
| core | static/methods-core.md |
| geometry | static/methods-geometry.md |
| semantic | static/methods-semantic.md |
| slam | static/methods-slam.md |
| dynamic | static/methods-dynamic.md |
| application | static/methods-application.md |
| all | ALL method fragments (methods-core.md through methods-application.md) |
Reference Load (when producing structured output)
- static/output-rules.md — Output format template and comparison rules
Optimization: For depth=quick, load only the relevant category fragment + core-stance. For depth=comprehensive, load all fragments plus output-rules.
Step 3 — Execute Comparison
After loading the required fragments:
- Apply the comparison dimensions from core-stance.md
- Use method data from the loaded category fragment(s)
- Follow output format from output-rules.md if producing tables/reports
- Observe all guardrails (no fabrication, flag uncertainty, load from disk not memory)
Cross-Skill Routing
This skill focuses on method-level comparison. For related workflows:
- Paper reading/analysis → 3dgs-paper-reader
- Code review/bug detection → 3dgs-code-reviewer
- Experiment design → 3dgs-experiment-planner
- Engineering deployment → 3dgs-engineering-guide
- Visualization/radar charts → 3dgs-visualizer
- CAD/Mesh integration → cad-mesh-3dgs
- Spatial intelligence/agent → 3dgs-spatial-agent
Red Lines
The following are categorical prohibitions. Violating any of these invalidates the output:
- No invented data: Never fabricate benchmark results, timing numbers, or method characteristics 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 comparisons: Do not reduce multi-dimensional trade-offs to a single "better/worse" judgment without context.
Related Skills
- 3dgs-paper-reader — Deep reading and analysis of 3DGS papers (use when you need detailed paper understanding, not just method comparison)
- 3dgs-code-reviewer — Bug detection and code review for 3DGS implementations (use when comparing implementation quality)
- 3dgs-experiment-planner — Experiment design for 3DGS research (use when comparison informs experiment design)
- 3dgs-engineering-guide — Production deployment of 3DGS (use when comparison targets deployment scenarios)
- 3dgs-visualizer — Publication-quality comparison visualizations (use when comparison needs radar charts or timeline plots)
- 3dgs-spatial-agent — Spatial intelligence and embodied reasoning with 3DGS
Guardrail: Do Not Apply From Memory
Do NOT try to apply the comparison logic, method data, 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, or technical features not present in the source data.
If you like it, please star this repo https://github.com/jaccen/Awesome-Gaussian-Skills
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
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.

