modal-research-gpu

GPU-accelerated scientific research on Modal — simulations, numerical methods, Monte Carlo, molecular dynamics, large-scale data processing. NOT for ML training or inference (use gpu-training or modal skills instead).

Quality

79%

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

Fix and improve this skill with Tessl

tessl review fix ./backend/cli/skills/cloud-compute/modal-research-gpu/SKILL.md

Quality

Content

82%Weight 40%Scale 1-5

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

A well-structured, highly actionable skill body dominated by executable code patterns and a clear cost-approval gate. The main gaps are minor verbosity in section intros and the absence of any bundle/reference files to offload domain-specific detail.

Suggestions

Add a post-execution verification step (spot-check small batch, then scale) explicitly into the batch/Monte Carlo patterns to create a validate-fix-retry feedback loop.

Move domain-specific deep dives (e.g. full OpenMM/LAMMPS configurations) into reference files and link to them one level deep to improve progressive disclosure.

Trim inline comments and section intros that restate library purposes Claude already knows (e.g. 'CuPy — NumPy API on GPU').

Dimension	Reasoning	Score
Conciseness	Largely lean with code-first patterns and minimal concept explanation, though a few section intros and inline comments restate things Claude already knows (e.g. 'CuPy — NumPy API on GPU').	4 / 5
Actionability	Provides multiple copy-paste-ready, executable code patterns covering the common cases (basic GPU compute, Monte Carlo fan-out, molecular dynamics, parameter sweeps, JAX PDEs) with concrete Modal APIs and GPU types.	5 / 5
Workflow Clarity	The Cost Management section gives a clear numbered pre-flight sequence plus an explicit CRITICAL approval gate (a validation checkpoint), but the destructive/batch patterns lack a verify-correctness feedback loop after execution.	4 / 5
Progressive Disclosure	Well-organized into clearly headed sections (When to Use, GPU Selection, patterns, Cost Management) that function as a navigable overview, but all content is inline with no one-level-deep references to separate detail files for the larger domain-specific topics.	4 / 5
	Total	17 / 20 Passed

Description

76%Weight 40%Scale 1-5

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, well-scoped description that names concrete scientific-computing domains and explicitly disambiguates from related ML/inference skills. It would benefit from an explicit 'Use when...' trigger clause and a few synonyms/file extensions to push completeness and trigger quality to the top level.

Suggestions

Add an explicit 'Use when...' clause naming the natural trigger phrases (e.g. 'Use when running GPU simulations, Monte Carlo studies, or molecular dynamics on Modal').

Include common synonyms and concrete file types (e.g. .pdb, .npz) to broaden trigger-term coverage.

Convert the enumerated capability areas into more concrete verb-led actions (e.g. 'run', 'parallelize', 'solve') to lift specificity.

Dimension	Reasoning	Score
Specificity	Lists several concrete capability areas ('simulations, numerical methods, Monte Carlo, molecular dynamics, large-scale data processing') with a clear domain, though they are category names rather than fully distinct concrete actions like 'extract' or 'fill'.	4 / 5
Completeness	Clearly answers 'what' (GPU-accelerated scientific research workloads enumerated) and provides strong 'when' guidance via the explicit NOT clause and skill-routing alternatives, but lacks an explicit 'Use when...' phrasing tying triggers together.	4 / 5
Trigger Term Quality	Contains natural terms users would say ('simulations', 'Monte Carlo', 'molecular dynamics', 'scientific research') plus the negative trigger guidance, though it lacks synonyms and common file-extension variants for some domains.	4 / 5
Distinctiveness Conflict Risk	Has a clear niche (GPU scientific computing) and explicitly disambiguates from adjacent skills via 'NOT for ML training or inference (use gpu-training or modal skills instead)', minimizing conflict risk.	5 / 5
	Total	17 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	15 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/cloud-compute/modal-research-gpu/SKILL.md
Commit: 3a6c3a9

Reviewed: 1 day 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.