Transformers
K-Dense-AI/scientific-agent-skills
Hugging Face Transformers for loading Hub models, running pipeline inference, text generation, and Trainer fine-tuning on NLP, vision, audio, and multimodal tasks.
Normalize a raw ONShape-exported YAM-family URDF into the aligned, world-referenced form used across i2rt robot models, as a two-stage pipeline around a human inspection checkpoint.
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdf --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .claude/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .claude/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdfType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdf --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .agents/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .agents/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .agents/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdf --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .cursor/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .cursor/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/i2rt-robotics/i2rt.git --path .agents/skills/transform-onshape-urdf--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdf --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .gemini/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .gemini/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdfInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .github/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .github/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .github/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install i2rt-robotics/i2rt transform-onshape-urdf --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/i2rt-robotics/i2rt.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/.agents/skills/transform-onshape-urdf .opencode/skills/transform-onshape-urdf && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "transform-onshape-urdf" agent skill from https://github.com/i2rt-robotics/i2rt/tree/main/.agents/skills/transform-onshape-urdf into .opencode/skills/transform-onshape-urdf/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "transform-onshape-urdf", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
transform-onshape-urdfNormalize a raw ONShape-exported YAM-family URDF into the aligned, world-referenced form used across i2rt robot models, as a two-stage pipeline around a human inspection checkpoint.
Transform Onshape Urdf is an agent skill from i2rt-robotics/i2rt. Normalize a raw ONShape-exported YAM-family URDF into the aligned, world-referenced form used across i2rt robot models, as a two-stage pipeline around a human inspection checkpoint. Use when importing a fresh ONShape URDF export (robot named "urdftopassembly", package:// mesh paths, a synthetic "root" link, dofjoint names, and duplicate 1 links wired by 分组/紧固 fixed joints); when removing a synthetic root link and baking its rotation into the base while keeping every downstream link in place; when applying a…
Its SKILL.md is about 2.4k tokens, which your agent loads only when the skill is triggered. The skill folder holds 5 other files, including scripts (for example `scripts/apply_urdf_heading.py`, `scripts/normalize_onshape_urdf.py` and `scripts/urdf_align_lib.py`).
The licence is MIT.
2 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 120c3c8. It shows what the files ask for, not the result of running them.
Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.
From allowed-tools in the SKILL.md frontmatter.
Ships 4 files in scripts/ (Python), which the agent can run.
Shell commands in SKILL.md call:
pythongitFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md. Its commands use git, which can reach the network depending on how they are called.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Transform Onshape Urdf loads about 2.4k tokens when it runs. Until then it costs about 217 tokens; SKILL.md has 1,052 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check found no risky patterns in SKILL.md.
Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
The full file from i2rt-robotics/i2rt at commit 120c3c8, republished under its MIT licence (© i2rt-robotics). 1,052 words, ~2,394 tokens.
.claude/skills/transform-onshape-urdf/SKILL.md (or your agent's skills folder). This skill also uses 4 other files; get the full folder from GitHub.Turns a raw ONShape YAM-family export into the clean form established by yam.urdf
(commit 8975c25). This was one script (transform_onshape_urdf.py); it is now two, split
around a mandatory visual-inspection checkpoint so the heading rotation -- a human judgment -- is
never applied as a default guess:
scripts/normalize_onshape_urdf.py — stage 1,
deterministic cleanup (meshes, joint names, remove root, bake R0). No heading.scripts/view_urdf.py — stage 2, the checkpoint: render the
normalized model in MuJoCo with world axes and ask the user which heading correction is needed.scripts/apply_urdf_heading.py — stage 3, apply the
user's rotation to the base (mesh + inertia together) and the arm.The shared math/XML helpers live in
scripts/urdf_align_lib.py.
The downstream URDF→MJCF alignment (regenerate the .xml, then re-sync gripper mounts) is owned by
.agents/skills/align-urdf-mjcf/SKILL.md. Treat the URDF as the
source of truth; regenerate the .xml (MJCF) from the fixed URDF.
<robot name="urdf_top_assembly">,
package://.../meshes/Name.stl paths, a root link above base, dof_joint0..N joints,
and duplicate link*_1 / 分组 / 紧固 structural joints.# Stage 1 — normalize (deterministic; edits in place):
python .agents/skills/transform-onshape-urdf/scripts/normalize_onshape_urdf.py <path/to/model.urdf> \
[--assets-dir assets] [--name NAME] [--dry-run] [--force]
# Stage 2 — inspect (required checkpoint):
python .agents/skills/transform-onshape-urdf/scripts/view_urdf.py <path/to/model.urdf> # or --screenshot out.png (headless)
# Stage 3 — apply the user's heading (previews by default; --apply writes):
python .agents/skills/transform-onshape-urdf/scripts/apply_urdf_heading.py <path/to/model.urdf> --axis z --deg <N> [--apply]normalize --name defaults to the URDF's parent dir name, skipping a vN version dir
(arm/yam/v1/yam.urdf -> yam). Pass it explicitly when the export sits somewhere unrelated.normalize --dry-run prints the verification report and the resulting base + joint1,
writing nothing. Always dry-run first.apply_urdf_heading writes only with --apply; without it, it previews. Different YAM arms need
different headings (each export's baked R0 differs) — inspect and confirm visually. Common case
is a yaw about world Z; yam_pro needed 0°.normalize_onshape_urdf.py does)RPY convention (URDF): R = Rz(yaw) @ Ry(pitch) @ Rx(roll).
dof_joint* joints. Drop
every other joint (分组/紧固/fastened duplicates) and every link not on that chain.
Rename the survivors to canonical names by stripping a trailing _<n>
(base_1→base, link2_1→link2, gripper_1→gripper, tip_right_1→tip_right), and
rename the base link to base whatever the export called it (ultra_base→base). The base is
identified structurally as the parent of dof_joint1, never by name — an export that names it
after the product would otherwise slip past step 3's child == "base" guard (reported as
"synthetic root already removed") and be rejected by every downstream stage, all of which require
the name base.assets/<file>, preferring an exact
case-insensitive match on disk, else the suffix-stripped lowercase name with step 1's link
renames applied to the stem (Gripper.stl→gripper.stl, link5_1.stl→link5.stl,
ultra_base.stl→base.stl) — a link and its mesh are always renamed together.R0. The synthetic root is the parent of the joint whose
child is base (in a raw export that is dof_joint0, root -> base). Capture its rotation
R0 = Rz(yaw)Ry(pitch)Rx(roll), delete the root link and that joint, and left-multiply R0
into the base visual origin, the base inertial origin (position + rpy), and the
base -> link1 joint origin. The inertia tensor components and every joint's local axis are
unchanged; joints 2..N are untouched, so the arm rotates rigidly. Rotating the base mesh and its
inertia by the same R0 keeps them consistent. Guard: if no joint has child base, the
root is already gone and this step is skipped — it never re-bakes and never mistakes a rootless
base for a synthetic root.dof_jointN → jointN; set the robot name.onshape-normalize: marker recording R0 and that a heading is still pending.No heading is applied here. That is stage 3, gated by the checkpoint.
apply_urdf_heading.py does)Construct the requested rotation R (--axis/--deg, or --rpy "r p y") and left-multiply it
into the base visual origin, the base inertial origin (the same R — mesh and inertia never
diverge), and the base -> link1 joint origin (rigidly rotating the whole arm). Because origins are
left-multiplied, this composes with stage 1's R0 as R @ R0 — identical to the old one-shot
transform's combined --heading-deg. Inertia components and downstream joints are untouched. An
onshape-heading: marker is appended. The op is not idempotent (re-running stacks another
rotation), so it previews by default and writes only with --apply.
Before doing anything, normalize_onshape_urdf.py classifies the file structurally:
is_raw: a root link, dof_joint* names, package:// paths, non-arm (分组/紧固)
joints, or *_<n> duplicate links. A non-raw file has nothing to normalize (each step is a
no-op), so it is skipped.has_marker: the onshape-normalize: comment is present.Decision: not raw → skip (already normalized). Raw + marker + no --force → skip. Raw and
(--force or no marker) → normalize, then write the marker. Every step is individually idempotent,
so re-running is safe.
Both stage scripts verify before writing and print a report:
base -> link1 are unchanged (the arm
is a pure rigid rotation).base root.R).After writing, additionally:
git diff -- <model.urdf> # base visual+inertial + joint1 rotated; root/dof_joint0 gone;
# joints renamed; mesh case fixed; other links byte-identicalThen inspect with python .agents/skills/transform-onshape-urdf/scripts/view_urdf.py <model.urdf> and confirm the arm assembles and
the base faces the intended forward direction — the required heading checkpoint from
align-urdf-mjcf. Finally regenerate the MJCF from the URDF (and re-sync gripper mounts — see
align-urdf-mjcf).
dof_joint*; structural duplicates use
分组/紧固; per-instance suffixes are _<n>; the base link is the parent of dof_joint1 and may
be named anything. Verified end-to-end on yam_pro and yam_ultra v2.--rpy). If a model needs a
different base orientation, pass a different --axis/--deg and confirm by rendering.R0 at 5 decimals (rpy="0 1.5708 ...", and 1.5708 - π/2 = 3.673e-6), so a
plain --deg heading leaves a ~0.0002° tilt in the base that every downstream frame inherits.
Since the heading left-multiplies (R_head @ R0), you can absorb it: pass the intended heading
composed with the correction via --rpy. For yam_ultra v2 (R0 = Ry(1.5708), heading z+180°)
that was --rpy "0 -3.673205103379957e-06 3.141592653589793" → the product is exactly
Rz(π)·Ry(π/2), giving base rpy 0 1.57079632679 3.14159265359 and joint1 xyz 0 0 0.0733.© i2rt-robotics, MIT. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 4 other files (scripts) in .agents/skills/transform-onshape-urdf of i2rt-robotics/i2rt.
Open the folder on GitHubat commit 120c3c8
Transform Onshape Urdf next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Transform Onshape Urdf this skilli2rt-robotics/i2rt | 165 | — | ~2.4k | Automated safety check: Pass | MIT | |
| TransformersK-Dense-AI/scientific-agent-skills | 48k | 1 repos | ~2.8k | Automated safety check: Notes | Apache-2.0 | |
| Export0xsline/OpenChatCut | 2.2k | — | ~2.2k | Automated safety check: Warn | AGPL-3.0 | |
| Export Download Debuggingnexu-io/open-design | 100k | — | ~725 | Automated safety check: Pass | Apache-2.0 | |
| Cost Exportruvnet/ruflo | 74k | 1 repos | ~687 | Automated safety check: Notes | MIT | |
| Hugging Face Transformers Usagedavila7/claude-code-templates | 32k | 12 repos | ~1.2k | Automated safety check: Pass | MIT |
K-Dense-AI/scientific-agent-skills
Hugging Face Transformers for loading Hub models, running pipeline inference, text generation, and Trainer fine-tuning on NLP, vision, audio, and multimodal tasks.
0xsline/OpenChatCut
A skill your agent uses when a OpenChatCut video editing or creation workflow needs export, render, download, share, final delivery, subtitle-file export, render choice, local-only asset handling…
nexu-io/open-design
Diagnose and fix browser, preview, or Electron export/download failures, especially image export issues involving Save As, Blob/Data URLs, the File System Access API, createWritable failures, and 0…
ruvnet/ruflo
Export cost-tracking telemetry in Prometheus textfile or webhook JSON formats — for external observability (Grafana, Datadog, custom dashboards)
davila7/claude-code-templates
Loads pre-trained Hugging Face Transformers models for text, vision and audio tasks, runs inference with pipelines and fine-tunes on custom datasets.
sickn33/agentic-awesome-skills
Use Transformers.js to run state-of-the-art machine learning models directly in JavaScript/TypeScript.
i2rt-robotics/i2rt
Align a MuJoCo MJCF robot model with a URDF source while preserving kinematics, geometry, masses, centers of mass, and inertia tensors.
Normalize a raw ONShape-exported YAM-family URDF into the aligned, world-referenced form used across i2rt robot models, as a two-stage pipeline around a human inspection checkpoint. Transform Onshape Urdf is an agent skill from i2rt-robotics/i2rt. Normalize a raw ONShape-exported YAM-family URDF into the aligned, world-referenced form used across i2rt robot models, as a two-stage pipeline around a human inspection checkpoint.
Transform Onshape Urdf fits situations like: importing a fresh ONShape URDF export (robot named urdftopassembly; package:// mesh paths; A synthetic root link; duplicate 1 links wired by 分组/紧固 fixed joints).
Run `npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a claude-code`. Or copy the skill folder (.agents/skills/transform-onshape-urdf in i2rt-robotics/i2rt) into .claude/skills/transform-onshape-urdf in your project. Claude Code loads it when a task matches its description.
Run `npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a codex`. Or copy the skill folder (.agents/skills/transform-onshape-urdf in i2rt-robotics/i2rt) into .agents/skills/transform-onshape-urdf in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add i2rt-robotics/i2rt --skill transform-onshape-urdf -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/transform-onshape-urdf, .gemini/skills/transform-onshape-urdf, .github/skills/transform-onshape-urdf and .opencode/skills/transform-onshape-urdf in your project.
Going by SKILL.md and its folder, Transform Onshape Urdf needs Python for the scripts in its folder and the command-line tools its instructions call (python and git). Our summary lists: Python 3.
SKILL.md contains no URLs. Its commands use git, which can reach the network depending on how they are called. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
Transform Onshape Urdf is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 2.4k tokens (SKILL.md is roughly 9.6k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
Skills that share tags, products or a category with Transform Onshape Urdf: Transformers (K-Dense-AI/scientific-agent-skills, 48k stars), Export (0xsline/OpenChatCut, 2.2k stars), Export Download Debugging (nexu-io/open-design, 100k stars) and Cost Export (ruvnet/ruflo, 74k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
i2rt-robotics (a GitHub organization) maintains it in i2rt-robotics/i2rt, which has 165 GitHub stars. The repository holds 2 skills in this directory. The repository was last updated on September 17, 2026.
Source: i2rt-robotics/i2rt on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.