compute-env-setup

Set up a compute environment on a remote provider so Claude Science jobs can run there. Covers direct SSH/conda hosts, Slurm clusters, container-via-bridge runners, and managed-API providers (Modal, GCP, RunPod). Use when standing up a new provider, porting an env to a different backend, adding a tool that needs its own software stack, or wiring weight caches. Triggers on "new compute provider", "set up env on", "port env to", "build GPU image", "weight cache", "compute_details", "conda env on the box", "apptainer on slurm".

Quality

82%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Quality

Content

65%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A rich, expert-level, highly actionable skill body with strong validation and diagnosis feedback loops. Its main weaknesses are some essayistic prose that could be tightened, an implicit top-level workflow ordering, and a broken reference path (remote-compute-modal/env-setup.md) alongside inline content that could be split out.

Suggestions

Trim the discursive flourishes (e.g. "The honest answer to which shape should I use..." and the opening philosophy paragraph) so every sentence is instructional.

Add a short numbered top-level procedure (read compute_details → recognize shape → build → populate weights → validate 3 levels → record in compute_details) to make the workflow ordering explicit alongside the existing section organization.

Fix or remove the missing `remote-compute-modal/env-setup.md` reference, and consider moving the spec-field table and the diagnosing-failures table into reference files so SKILL.md reads as an overview with one-level-deep pointers.

Dimension	Reasoning	Score
Conciseness	The body is dense and assumes Claude's competence (no explanations of conda/containers/Slurm), but essayistic flourishes like "The honest answer to which shape should I use is usually the one this provider already is. You're rarely choosing; you're recognising" and the discursive opening overview add length without instruction, so it is mostly efficient rather than lean.	2 / 3
Actionability	Concrete, copy-pasteable commands appear throughout — `conda create -n <name> python=<X>`, `apptainer pull <name>.sif docker://<ref>`, `module use $HOME/modulefiles`, `du -sh $CACHE_VAR`, `OMP/MKL/OPENBLAS_NUM_THREADS=<tier.cpus>` — plus a symptom→fix diagnosis table and a worked compute_details ledger block, matching the fully-executable anchor.	3 / 3
Workflow Clarity	Strong validation checkpoints and feedback loops exist (the three validation levels, the diagnosis table as error-recovery, the "after any env rebuild or doc edit" guidance), but the overall build flow (read compute_details → recognize shape → build → populate weights → validate → record) is never rendered as one explicit numbered sequence, so checkpoints are present but ordering is implicit.	2 / 3
Progressive Disclosure	Sections are well organized and `references/envs_reference.md` is signaled and exists, but `remote-compute-modal/env-setup.md` is referenced and does not exist in the bundle, and large blocks (the spec table, the diagnosis table) stay inline rather than in separate reference files, fitting the some-structure-but-inline anchor.	2 / 3
	Total	9 / 12 Passed

Description

100%

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong description: concrete actions, an explicit Use-when clause, natural trigger phrases, and a clearly bounded niche in third-person voice. There is little to improve.

Dimension	Reasoning	Score
Specificity	Lists multiple concrete actions — "Set up a compute environment", "porting an env to a different backend", "adding a tool that needs its own software stack", "wiring weight caches" — and enumerates concrete provider types (SSH/conda, Slurm, bridge runners, Modal/GCP/RunPod), matching the multiple-specific-actions anchor.	3 / 3
Completeness	Explicitly answers both what ("Set up a compute environment on a remote provider so Claude Science jobs can run there") and when ("Use when standing up a new provider, porting an env to a different backend...") with an explicit Use-when clause, matching the top anchor; not below 3 because the when-guidance is explicit, not implied.	3 / 3
Trigger Term Quality	Trigger phrases like "set up env on", "port env to", "build GPU image", and "weight cache" are natural phrasings a user would actually say; even the jargon ones ("conda env on the box", "apptainer on slurm") reflect genuine user requests, giving good coverage.	3 / 3
Distinctiveness Conflict Risk	It carves a clear niche (remote compute-env setup for Claude Science jobs) with distinct, domain-specific triggers unlikely to fire for unrelated skills, matching the clear-niche anchor.	3 / 3
	Total	12 / 12 Passed

Validation

100%

Warnings & errors only

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: UnicomAI/wanwu
Path: configs/microservice/bff-service/configs/agent-skills/claude-science/compute-env-setup/SKILL.md
Commit: f618458

Reviewed: about 13 hours ago

Table of Contents

Discovery Implementation Validation

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.