blender-pipeline-automation

Before you start: read this whole install prompt, tell me in a few lines what this skill does and what installing it involves, and wait for my go-ahead before you do anything else. Prerequisites: Blender 5.2 LTS or newer is required to USE this skill (installing it is only copying files, but every use runs code inside Blender). No account, no API key, no paid service, and nothing is sent anywhere by the skill's scripts. Two optional patterns are documented in SKILL.md and are NOT done by this install and never run automatically: using Blender as a Python module via pip install bpy==5.2.* (a roughly 300 MB download from PyPI, needs Python 3.13), and a CI recipe that downloads the Blender Linux tarball from download.blender.org. SKILL.md also says its execution model lives in the separate blender-core skill (its run_in_blender.py launcher and its references/bpy-patterns.md), which this listing does NOT install; the batch script in this listing runs directly inside Blender headless without that launcher. This listing installs only the blender-pipeline-automation skill: the 21 other blender-* skills in the same repository are NOT installed, even though SKILL.md gives examples that run other skills' scripts. Install "blender-pipeline-automation" for me. It is the batch and pipeline skill for Blender 5.2: headless CLI use, batch conversion, inventory reports and running another script over whole folders of .blend, FBX, glTF, OBJ, STL, PLY, USD and Alembic files, with sharding across parallel processes, library linking and path management, and render-farm and CI patterns. Repository: https://github.com/EL4CTEO/blender-skills (the skill is the whole skills/blender-pipeline-automation/ folder at the pinned commit: SKILL.md, scripts/_bpyx.py, scripts/batch_process.py) Pinned version: install the code at commit 71d031dab8bc6729504d11a78cbe91945401045f exactly. Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Clone https://github.com/EL4CTEO/blender-skills into a temporary folder, check out commit 71d031dab8bc6729504d11a78cbe91945401045f, and copy the skills/blender-pipeline-automation/ folder from that checkout. Then compare every file in the folder with the fingerprints recorded for this listing on 2026-10-19: [FINGERPRINTS] - SKILL.md: SHA-256 590e1159205c68af5c338425456c3ff86837cdb8202e34f2d40a018e96ec79fa - scripts/_bpyx.py: SHA-256 a3339ba8a678469e8ed29510e32a7331e0235202191ca4edb977af4720819b4b - scripts/batch_process.py: SHA-256 baedf17711a939981f3796879e1f32be3203c197c41d9dc3fad8a104b82bfe8e If a file does not match, it changed since the listing was prepared: do not install it yet, tell me what you found and let me decide. 2. Read SKILL.md in full and tell me in plain words: what the skill does when used, which services it connects to, whether it reads or sends any personal information or keys, and anything that seems unusual. Stop on any red flag and tell me before going further. Confirm these points from the files: - The skill folder holds three files: SKILL.md (4,899 bytes), and two Python scripts under scripts/ (scripts/_bpyx.py, 4,503 bytes; scripts/batch_process.py, 6,833 bytes). There is no references/ folder in this skill. - Nothing is sent anywhere during installation itself; the only download at install time is the skill folder from GitHub at the pinned commit. Neither script contains any network code (no HTTP, no sockets, no uploads), no telemetry, no account and no API key, and neither script reads any environment variable or calls subprocess. The only hosts named in SKILL.md are download.blender.org (the optional CI tarball) and PyPI behind the optional pip install bpy pattern described in the prerequisites; neither runs at install time or automatically at use time. The environment variable names in SKILL.md (BLENDER_USER_SCRIPTS, BLENDER_USER_EXTENSIONS, BLENDER_SYSTEM_PYTHON and OCIO) are Blender's own documented variables, not values the skill's scripts read. - scripts/batch_process.py runs inside Blender and processes every matching file from a folder, a glob or a list. In report mode it returns an inventory as JSON per file (object kinds and counts, triangles, materials, images, missing image paths and actions). In convert mode it imports each file and exports it to glb, fbx, obj, stl, usd, usdz or blend, writing to the --out folder given; if no --out is given, the converted file is written next to its source file, so I should always pass --out. In script mode it loads the Python file I name with --script (using importlib, which executes that file) and calls its run() with the JSON --kwargs I provide for each input file, optionally saving a .blend copy per file to --save-to. Files can be split across parallel Blender processes with --shard i/n. A failure on one file does not stop the batch, and the summary lists the failures. Loading a file replaces the current Blender session content: .blend files are opened with open_mainfile, and other formats start from read_factory_settings(use_empty=True) before import, so using it through MCP on an open file replaces that open file, as its own docstring warns. - scripts/_bpyx.py is the shared helper batch_process.py imports (argument parsing after "--", JSON output on a marked line, CLI wrapper, object and node helpers); it is a synced copy of the repository's shared/_bpyx.py, byte-identical at the pinned commit. Its CLI wrapper can save a .blend only when a script is run with an explicit save path. - At use time, everything runs locally in my own Blender, headless (blender -b -P scripts/batch_process.py -- ...), on folders and files I point you at; no use is launched by this install. SKILL.md's own robustness rules apply: never write into the source folder by default (use a separate --out or --save-to, and save with copy=True), make scripts idempotent, log JSON per file, and pin Blender 5.2 LTS for a project. - The scope is batch and pipeline work only. The execution foundation SKILL.md points to (blender-core, its run_in_blender.py launcher and bpy patterns reference) and the specialised scripts its examples run per file (for example blender-asset-qa and blender-game-optimization scripts) are other blender-* skills, which this listing does not install; the repository's own installer and Claude Code plugin install all 22 skills at once and are not the route used here. 3. When I say "yes, install", install it as a skill: copy the whole blender-pipeline-automation folder into the agent's skills directory, in a folder named "blender-pipeline-automation", keeping all three files with their scripts/ subfolder. 4. Verify with no network calls: frontmatter valid, files in place. 5. Report what was installed, where, and what I still need to do myself (for example: install Blender 5.2 LTS or newer myself if I do not have it, install the separate blender-core skill myself if I want its launcher and patterns, install any other blender-* skill separately if I want its scripts for batch script mode, and decide myself whether I ever use the optional bpy pip module or the CI tarball pattern; using the skill is a separate step that only starts when I ask for it). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.