flow-nexus-neural

Train and deploy neural networks in distributed E2B sandboxes with Flow Nexus

2.63x

Quality

51%

Does it follow best practices?

Impact

95%

2.63x

Average score across 3 eval scenarios

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./.claude/skills/flow-nexus-neural/SKILL.md

Quality

Content

50%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.

The body is highly actionable, with concrete executable MCP examples for every capability, but it is a monolithic, heavily duplicated reference document. Duplicated examples inflate token cost, and multi-step cluster workflows lack validation checkpoints, placing it at the rubric midpoint overall.

Suggestions

Deduplicate the LSTM and transformer configs (each appears 2-3 times) and move per-tool API details and architecture patterns into reference files under references/ so SKILL.md stays a lean overview.

Add explicit validation checkpoints to cluster workflows, e.g. poll neural_cluster_status and verify node status/loss before starting distributed training, and confirm status before neural_cluster_terminate.

Replace pseudocode placeholders ("generator_layers: [...]") with real layer definitions and show how to obtain the real user_id instead of "your_user_id".

Dimension	Reasoning	Score
Conciseness	The ~727-line body repeats the same LSTM layer config three times (training example, Time Series Forecasting use case, and Architecture Patterns) and the transformer config likewise, with sample responses padding nearly every tool call. This is noticeably verbose with several unnecessary duplicated sections, matching anchor 2 rather than the mostly-efficient anchor 3.	2 / 5
Actionability	Concrete MCP calls with full parameter objects and sample responses (e.g. mcp__flow-nexus__neural_train({...}), neural_cluster_init({...}), neural_cluster_status) make the guidance mostly executable. It falls short of 5 because the GAN pattern uses pseudocode ("generator_layers: [...]") and placeholders like "your_user_id" leave key details unresolved.	4 / 5
Workflow Clarity	Multi-step sequences exist (cluster init → node deploy → connect → train → monitor → terminate) but validation checkpoints are absent: destructive and batch operations like cluster_terminate and distributed training jobs have no verify-before-proceed steps, and the rubric caps workflow clarity at 3 in that case. Monitoring calls exist, but no feedback loop tells Claude what to check or when to abort.	3 / 5
Progressive Disclosure	There are no bundle files at all — the entire tool API reference, architecture patterns, use cases, and troubleshooting are inlined in one monolithic 727-line SKILL.md. Section headers give real structure (better than anchor 2's no-headers example), but content that clearly belongs in reference files (per-tool API details, architecture patterns) is inline, matching anchor 3.	3 / 5
	Total	12 / 20 Passed

Description

53%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.

The description states a clear what with a distinct Flow Nexus/E2B niche, but it is a single bare sentence: it names only two actions and provides no when-to-use guidance or the fuller capability list. It sits squarely at the rubric's midpoint.

Suggestions

Add an explicit trigger clause, e.g. "Use when training or deploying neural networks, running distributed/federated training, or browsing the model template marketplace."

List the additional concrete capabilities (inference, template marketplace, distributed clusters, benchmarking) so specificity moves from 1-2 actions toward comprehensive coverage.

Include natural synonyms such as "deep learning", "model training", and "fine-tuning" so trigger terms match what users actually say.

Dimension	Reasoning	Score
Specificity	"Train and deploy neural networks in distributed E2B sandboxes" names the domain and exactly two concrete actions (train, deploy), matching the 1-2-concrete-actions anchor. It is not a 4 because several real capabilities (inference, template marketplace, distributed clusters, benchmarking) are absent from the description.	3 / 5
Completeness	The "what" is clear (train and deploy neural networks in E2B sandboxes), but there is no "Use when..." clause or any equivalent trigger guidance; the rubric explicitly caps completeness at 3 in that case.	3 / 5
Trigger Term Quality	"train", "deploy", "neural networks", "distributed", and "E2B sandboxes" are relevant keywords, but common variations users would actually say ("deep learning", "model training", "fine-tune", "GPU training") are missing, matching the some-keywords-but-missing-synonyms anchor.	3 / 5
Distinctiveness Conflict Risk	"Flow Nexus" and "distributed E2B sandboxes" carve a clear niche with distinct triggers, leaving only minor overlap risk with generic ML/model-training skills; not a 5 because "train and deploy neural networks" alone is broad enough to collide with similar ML skills.	4 / 5
	Total	13 / 20 Passed

Validation

87%

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 — 14 / 16 Passed

Validation for skill structure

Criteria	Description	Result
skill_md_line_count	SKILL.md is long (757 lines); consider splitting into references/ and linking	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: ruvnet/RuVector
Path: .claude/skills/flow-nexus-neural/SKILL.md
Commit: 4b59352

Reviewed: 3 days 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.