Menu

Individual skill

setup-colabfold-remote-tool

0upvotes

Set up, launch, validate, and troubleshoot the ColabFold (AlphaFold2-Multimer) ToolUniverse remote tool and optionally relay it through ToolUniverse Connect. Use when deploying or auditing this implementation.

skills/setup-colabfold-remote-tool/SKILL.md

Read pinned source

Source commit 5598a9fc97e32964fe9b703818e7d0747142f179

What was checked

Matches the observed repository HEAD
Yes
Scanner rules are current
Yes
Repository scan complete
No
Skill scan complete
No
Observed behavior
Network or broad shell
Last checked
Nov 9, 2026

These are observations at the recorded commit, not a guarantee of safety. Review the source before running it.

Install this skill

Choose an agent in your Stack or copy a command below. Commands fetch the pinned commit and copy only this skill directory. Existing destinations are refused. Copying a command does not install anything.

An install command is unavailable: current rules, complete repository and skill scans, an eligible behavior level and a permitted license are required.

Repository context

Repository stars (not skill ratings)
1,780
Repository last updated
Nov 8, 2026
License
Apache-2.0

Observed agent integration: No specific integration evidence. Standard SKILL.md installation does not imply verified compatibility.

Files and excerpt

File metadata and permitted, redacted excerpts come from the recorded scan. Source links are pinned to that commit.

Browse files
Read excerpt

# Set up ColabFold as a remote tool

> Validation status (2026-10-01): ColabFold 1.6.1 on AlphaFold2-Multimer v3 ran on the NVIDIA GB10 (aarch64, cc 12.1, driver 580.173.02) under the optimization kit's `off` mode, on jax/jaxlib 0.5.3. **It takes a precomputed MSA and contacts no MMseqs2 server** — that is upstream's own branch (`run()` is `if a3m_lines is None: get_msa_and_templates(...) else: unserialize_msa(...)`), and `run_mmseqs2` is imported inside the branch this provider never takes. **Stock ColabFold cannot fold on this card at all**: the forward pass aborts with `Unsupported conversion from bf16 to f16 / LLVM ERROR: Unsupported rounding mode for conversion`, an uncatchable `report_fatal_error` from LLVM's NVPTX instruction selector, because ColabFold keeps its two recycle-state tensors in float16 while the trunk is bf16. Upstream's own `use_bfloat16=False` clears it, and that parameter has no command-line flag — `colabfold_batch` exits 134 (SIGABRT) on this hardware having written logs and no structure. So this wrapper is not a faster route to the model here; it is the only one. 17 requests, 17 ranked structures, 0 failures. Ladder 4.0 GiB flat at 100 / 200 / 400 residues (declared at 400); residency 1.16x warm, because upstream persists the expensive part. Public publication, cross-user isolation, concurrency, recovery and scientific accuracy are unvalidated; keep this deployment private.

## Prerequisites

- Run from the ToolUniverse repository root on Linux with Python 3.12.3.
- A CUDA GPU is required. This provider holds its model resident between requests.
- The model environment needs **Python 3.11** (`uv venv --seed --managed-python --python 3.11 ~/colabfold-venv`).
- Keep provider data, weights, caches, and credentials outside Git.
- Bind to loopback. A non-loopback bind requires TOOLUNIVERSE_API_TOKEN; never put it in arguments or results.

Run the standard-library contract check before downloading large dependencies:
Back to top