Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add IvanYangYangXi/artclaw_bridge --skill scene-vision-analyzergit clone --depth 1 https://github.com/IvanYangYangXi/artclaw_bridgeWrote 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/ivanyangyangxi/artclaw_bridge/scene-vision-analyzer)<a href="https://agentmods.dev/skills/ivanyangyangxi/artclaw_bridge/scene-vision-analyzer"><img src="https://agentmods.dev/badge/skills/ivanyangyangxi/artclaw_bridge/scene-vision-analyzer/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/ivanyangyangxi/artclaw_bridge/scene-vision-analyzer"><img src="https://agentmods.dev/badge/skills/ivanyangyangxi/artclaw_bridge/scene-vision-analyzer.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 | $0.00133 | $0.01058 |
| Opus 5 | $0.00067 | $0.00529 |
| Sonnet 5 | $0.00027 | $0.00212 |
| Haiku 4.5 | $0.00013 | $0.00106 |
Grade A, and why
scene-vision-analyzer 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 12d 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 — 101 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Scene Vision Analyzer
Analyze game scene concept art into structured JSON — the data foundation for 2D→3D reconstruction pipeline.
Workflow
Step 1: Prepare
- Load the image (user provides path or URL)
- Read
references/output-schema.jsonfor the full JSON schema - Read
references/analysis-prompts.mdfor prompt templates
Step 2: Run Analysis (Round 1 — Required)
Send the image to a multimodal AI with:
- System prompt: from
analysis-prompts.md§ System Prompt - User prompt: from
analysis-prompts.md§ User Prompt, with{schema}replaced byoutput-schema.jsoncontent - Image: attached as vision input
Parse the AI response as JSON. If the response is wrapped in markdown code blocks, strip them.
Step 3: Refinement (Round 2 — Recommended)
Send the Round 1 result back with the refinement prompt from analysis-prompts.md § 第二轮细化 Prompt.
Focus: missed objects, bbox precision, ground_contact accuracy.
Step 4: Consistency Validation (Round 3 — Optional)
Send the Round 2 result with the validation prompt from analysis-prompts.md § 第三轮一致性校验 Prompt.
Focus: size consistency, depth consistency, occlusion logic, shadow direction, group coherence.
Step 5: Save Results
Use scripts/scene_analyze.py helper functions:
from scene_analyze import save_result, generate_summary
# save_result(output_dir, json_string) → saves analysis_result.json + analysis_summary.txt
Or save manually:
analysis_result.json— full structured resultanalysis_summary.txt— human-readable summary
Default output dir: <image_dir>/scene_analysis_<timestamp>/
Quick Reference: Key Schema Fields
| Field | Purpose | Used by downstream |
|---|---|---|
objects[].ground_contact_pct |
Where object touches ground | Step 3 (perspective → 3D coords) |
objects[].estimated_size_m |
Real-world size in meters | Step 3 (scale calibration) |
objects[].bbox_pct |
2D bounding box (%) | Step 2 (annotation drawing) |
camera.pitch_angle_deg/yaw_angle_deg |
View angles | Step 3 (camera matrix) |
camera.horizon_position_pct |
Horizon line position | Step 3 (vanishing point) |
spatial_references.perspective_lines |
Depth cue lines | Step 3 (homography) |
spatial_references.vanishing_points_pct |
Convergence points | Step 3 (focal length est.) |
ground_plane.polygon_pct |
Ground area outline | Step 3 (ground plane fit) |
groups |
Repeated object patterns | Step 5 (batch placement) |
What ships with it
4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 12d ago First seen · 101 lines · 133 tokens per session scan A 69402bd301e4
scene-vision-analyzer is a skill published in the GitHub repository IvanYangYangXi/artclaw_bridge (35 stars, last pushed 4mo ago), licensed MIT. It adds 133 tokens to every session and 1,058 once invoked, about $0.0007 per session on Opus 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 2026-08-30.
Other skills, from other repositories
usd-tools
Infrastructure skill — low-level OpenUSD scene inspection and validation: read layer stacks, traverse prims, validate USD schemas. Use when working directly with raw USD files (usda, usdc, usdz) or verifying USD compliance. Not for Maya-specific USD export — use maya-pipelineexportusd for that. Not for full DCC…
spatial-interchange
Plan deterministic coordinate-axis and unit conversions between DCCs, engines, and interchange formats from explicit right/up/forward axes and meters-per-unit values.
game-level-layout
Game level layout review helpers for editor contexts.
dcc-mcp-core
Foundation library for the DCC Model Context Protocol (MCP) ecosystem. Provides Rust-powered action management, skills system, IPC transport, MCP Streamable HTTP server (2025-03-26 spec, with 2025-06-18 and 2025-11-25 awareness), sandbox security, shared memory, screen capture, USD scene support, and telemetry for…
dcc-cua
A routing guide for controlling application interfaces through the project’s DCC-CUA system. DCC applications are digital-content tools such as Maya, Blender, and Houdini.
marketplace-publish-extension
Infrastructure skill — publish (register/update) an extension package to a marketplace catalog. Reads the extension's SKILL.md frontmatter, constructs a CatalogEntry, and upserts it into the target marketplace.json. Optionally commits and pushes when the catalog source is a git repository. Use after scaffolding an…