propose_keypoints
100 / 100 · Low risk
Fine to try.
ReKep-style semantic keypoint proposer. For a named object: SAM-mask it, extract DINOv2 patch features inside the mask, k-means(cosine) cluster into K groups. Returns K cluster-centroid pixels. Each centroid lands on an anatomically meaningful sub-region (head vs handle, strike face vs claw, bottle neck vs body). MUCH better than uniform-stride sampling when the object has color/texture variation that maps to semantic parts. Pair with label_points_grid or unproject_pixel to get world XYZ for fi…
What this skill can do
We found nothing in the files we read that matches our rules. That does not prove the skill is safe: see the limits of static scanning.
Install
Paste this into your terminal. It downloads exactly the version we scanned, not whatever is newest.
if [ -e ~/.agents/skills/'propose-keypoints' ] || [ -L ~/.agents/skills/'propose-keypoints' ]; then echo 'A skill folder named propose-keypoints already exists — not replacing it' >&2; false; else d=$(mktemp -d) && git clone --quiet --filter=blob:none --no-checkout 'https://github.com/nssmd/RoboRSI' "$d" && git -C "$d" checkout --quiet '90c8569705bf4004a4f79e34e650c87ab7f075d4' -- 'roborsi/embodied/skills/base/propose_keypoints/robotwin' && mkdir -p ~/.agents/skills && cp -R "$d"/'roborsi/embodied/skills/base/propose_keypoints/robotwin' ~/.agents/skills/'propose-keypoints' && rm -rf "$d"; fi
Codex reads ~/.agents/skills directly. For Claude Code, Cursor, Copilot, Gemini CLI, Antigravity or OpenCode, also link it. For Claude Code: ln -s ~/.agents/skills/'propose-keypoints' ~/.claude/skills/'propose-keypoints' (other agents: ~/.cursor/skills, ~/.copilot/skills, ~/.gemini/skills, ~/.gemini/config/skills, ~/.config/opencode/skills). Or install through the finder skill, which does both.
Badge for your README
[](https://zalavijaysinh.github.io/skilltrust/repos/nssmd/RoboRSI/)
Score history
| Date | Version | Score |
|---|---|---|
| 2026-11-13 | 90c8569 | 100 |