"Explains VLA-Adapter Prismatic/VLA APIs, robot constants,
Scanned 9/9/2026
Install to Claude Code
npx -y skills add VectorSpaceLab/AREX-Skill --skill package-apis --agent claude-codeInstalls into .claude/skills of the current project.
Are you the author of Package Apis?
Add the live security badge to your README — it updates automatically with every re-scan.
[](https://www.skillsdirectory.com/skills/vectorspacelab-package-apis-arex-skill)More formats (shields.io, HTML) on the badges page.
---
name: package-apis
description: "Explains VLA-Adapter Prismatic/VLA APIs, robot constants,
checkpoint layouts, LoRA merge, and HF conversion utilities."
disable-model-invocation: true
metadata:
disco-role: operating
license: MIT
---
# Package Layout/API Reference (external-checkout adapter)
This sub-skill documents the native package layout and public API surface; it
does not implement or vendor that package. The generated skill directory
contains only this documentation, references, and the read-only checkpoint
layout validator. The native `prismatic/`, `vla-scripts/`,
`experiments/robot/`, and conversion scripts remain in the separate checkout.
Before inspecting any native source or installed package, set the absolute root
and enter it:
```bash
export VLA_ADAPTER_REPO_ROOT=/abs/path/to/VLA-Adapter
cd "$VLA_ADAPTER_REPO_ROOT"
python -m pip install -e "$VLA_ADAPTER_REPO_ROOT"
```
External prerequisites include the native checkout, its base Python package
dependencies, a compatible checkpoint, and (only for conversion or model
execution) the matching CUDA/Hugging Face stack. This sub-skill is not
self-contained and does not provide `load`, `load_vla`, action-head,
projector, LoRA-merge, or HF-conversion implementations. It does not load
weights or execute a model.
## Scope
Use this sub-skill to read [references/api-reference.md](references/api-reference.md)
and [references/prismatic-apis-and-conversion.md](references/prismatic-apis-and-conversion.md),
map native entrypoints and checkpoint files, or run the local layout validator.
Treat the API signatures and conversion inventory as reference facts, not
copyable implementations.
Run the skill-local validator by absolute skill path:
```bash
python "$VLA_ADAPTER_SKILL_ROOT/sub-skills/package-apis/scripts/check_checkpoint_layout.py" \
--checkpoint "$VLA_ADAPTER_REPO_ROOT/outputs/CHECKPOINT" --help
```
The validator only inspects local files. For native operations, use the native
entrypoints from `<absolute-repo-root>` after separately confirming their
prerequisites; this adapter never invokes them.
Is this your skill, or is something wrong with this listing? Request removal or report an issue. Author removals are honored within 72 hours.
No comments yet. Be the first to comment!