Menu

Individual skill

setup-e1-remote-tool

0upvotes

Set up, launch, validate, and troubleshoot the Profluent E1 ToolUniverse remote tool and optionally relay it through ToolUniverse Connect. Use when deploying or auditing this implementation.

skills/setup-e1-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 Profluent E1 as a remote tool

> Validation status (2026-10-01): Profluent E1 600m ran on the NVIDIA GB10 (aarch64, cc 12.1) with the optimization kit **OFF** -- stock E1, and the only reachable configuration here rather than a fallback: flash-attn publishes no aarch64 build for this torch, and the kit's accelerator proof demands the literal string `2.8.0+cu128`, which no aarch64 environment can satisfy because no CUDA build of torch 2.8.0 exists for this architecture at all (only `+cu129`). Verified on FOUR CONSECUTIVE IDENTICAL REQUESTS plus a mixed-shape sequence in one held worker: scoring in 0.075-0.080 s per request after a 12 s first call, peak reserved 4.30 GiB at 100 residues rising only to 4.84 at 800. That is the flattest ladder of any provider here -- 1.13x over 8x the length -- because E1 caps its batch at 65,536 tokens, so doubling the parent halves the batch. Public publication, cross-user isolation, concurrency, recovery and scoring quality are untested; 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; CPU is not a target.
- 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:

~~~bash
Back to top