Scene Loader
- 1 installs
- Updated May 7, 2026
- evenpang666/defense_ei
Idempotently loads a MuJoCo scene and saves a front-camera image before supervisor planning, skipping the render when a fresh image already exists.
About
Guarantees a usable current_view.png at the orchestrator's path by calling a scene-loader script at most once per scene, verifying the ready signal. A developer uses it as a bootstrap step before agent planning in a MuJoCo-based robotics pipeline.
- Zero-tool-call skip when a fresh image is already on disk
- Single-shot load with strict ok/signal verification and no retries
Scene Loader by the numbers
- 1 all-time installs (skills.sh)
- Ranked #14,098 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
- Data as of Jul 8, 2026 (Skillselion catalog sync)
npx skills add https://github.com/evenpang666/defense_ei --skill scene-loaderAdd your badge
Show developers this skill is listed on Skillselion. Paste this into your README.
| Installs | 1 |
|---|---|
| Last updated | May 7, 2026 |
| Repository | evenpang666/defense_ei ↗ |
What it does
Idempotently loads a MuJoCo scene and saves a front-camera image before supervisor planning, skipping the render when a fresh image already exists.
Files
Scene Loader Bootstrap Skill
Guarantee a usable front-camera image at the path the orchestrator passes (default logs/current_view.png) before any planning step. Optimize for zero tool calls when the image already exists and is fresh.
---
Tool
scene_loader_script(scene_json_path, current_view_path, randomize_texture=False)— loads the MuJoCo scene and saves the front-camera image. Returns JSON withok,signal,scene_json_path,current_view_path.
---
Decision flow
1. Skip-if-fresh: if the user message states current_view.png exists, or the orchestrator already attached the image, do not call the tool. Treat the image as ready. 2. Single-shot load: otherwise call scene_loader_script(scene_json_path, current_view_path, randomize_texture=False) exactly once.
- Use
scene_json_pathfrom the user message (defaultchemistry.json). - Use
current_view_pathfrom the user message (defaultlogs/current_view.png).
3. Verify response: require ok=true and signal="CURRENT_VIEW_READY". The returned current_view_path is now the canonical image for planning. 4. Failure handling: if ok=false, return the raw JSON error verbatim and stop. Do not retry, do not invent a plan.
---
Iteration budget
scene_loader_scriptcalls per session: ≤ 1 in the happy path, ≤ 2 only if the orchestrator explicitly requests a re-render with a newscene_json_path.
---
What NOT to do
- Do not call this tool on every turn — once the image exists for the active scene, treat it as immutable for the rest of the session.
- Do not enable
randomize_textureunless the user message explicitly asks for randomized textures. - Do not produce planning output from this skill; that belongs to
supervisor-contract.
Related skills
Setup Matt Pocock SkillsScaffold the per-repo configuration that Matt Pocock’s engineering agent skills rely on so they understand the issue tracker, triage labels, and domain documentation la495k196k
Lark AppsConnect Claude, Cursor or custom agents directly to Lark (Feishu) for messaging, document automation, approval workflows and enterprise data access.402k
CavemanSlash token usage by roughly 75% while keeping every technical detail intact when working with Claude Code, Cursor or similar agents.391k94.7k
Lark Skill MakerQuickly turn any Lark/Feishu OpenAPI call or multi-step workflow into a reusable agent skill with its own SKILL.md.386k16k
Running Claude Code Via Litellm CopilotRun Claude Code at a fraction of the cost by routing requests through LiteLLM to the GitHub Copilot Chat API.270k72
Codex PetGenerate a complete Codex Pet spritesheet and metadata from one reference image without needing an OpenAI key or Codex Pro.246k10