flow-nexus-neural
Train and deploy neural networks in distributed E2B sandboxes with Flow Nexus
57
37%
Does it follow best practices?
Impact
96%
7.38xAverage score across 3 eval scenarios
Advisory
Suggest reviewing before use
Optimize this skill with Tessl
npx tessl skill review --optimize ./ai-ml/flow-nexus-neural/SKILL.mdQuality
Discovery
32%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 identifies a specific domain (neural network training/deployment) and mentions distinctive tooling (E2B sandboxes, Flow Nexus), but it is too terse and lacks a 'Use when...' clause, making it difficult for Claude to know when to select this skill. It also misses common user-facing trigger terms like 'machine learning', 'deep learning', or 'model training'.
Suggestions
Add an explicit 'Use when...' clause, e.g., 'Use when the user asks about training, fine-tuning, or deploying machine learning models in sandboxed environments.'
Include natural trigger terms users would say, such as 'machine learning', 'deep learning', 'model training', 'ML deployment', 'GPU training'.
Expand the list of specific capabilities beyond just 'train and deploy', e.g., 'configure distributed training jobs, monitor training progress, manage model checkpoints, deploy inference endpoints'.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Names the domain (neural networks, distributed sandboxes) and two actions (train and deploy), but lacks comprehensive detail about specific capabilities beyond those two verbs. | 2 / 3 |
Completeness | Describes what it does (train and deploy neural networks) but completely lacks a 'Use when...' clause or any explicit trigger guidance for when Claude should select this skill. Per rubric guidelines, missing 'Use when' caps completeness at 2, and the 'what' is also thin, so this scores a 1. | 1 / 3 |
Trigger Term Quality | Includes relevant terms like 'neural networks', 'distributed', and 'deploy', but 'E2B sandboxes' and 'Flow Nexus' are product-specific jargon that users may not naturally say. Missing common variations like 'deep learning', 'model training', 'ML', 'machine learning'. | 2 / 3 |
Distinctiveness Conflict Risk | The mention of 'E2B sandboxes' and 'Flow Nexus' adds some distinctiveness, but 'train and deploy neural networks' is broad enough to overlap with general ML/AI skills. The product-specific terms help but the core action is generic. | 2 / 3 |
Total | 7 / 12 Passed |
Implementation
42%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
The skill is highly actionable with concrete, executable MCP tool calls and realistic examples covering the full API surface. However, it is extremely verbose and repetitive — architecture examples appear multiple times, use cases duplicate earlier content, and the entire document is a monolithic block that should be split across multiple files. Workflow clarity suffers from missing validation checkpoints in multi-step distributed training operations.
Suggestions
Extract Architecture Patterns, Common Use Cases, and detailed API examples into separate referenced files (e.g., ARCHITECTURES.md, EXAMPLES.md) and keep SKILL.md as a concise overview with links.
Remove duplicate content — the architecture configs in 'Architecture Patterns' repeat what's already shown in the training examples above.
Add explicit validation checkpoints to the distributed training workflow (e.g., 'Verify cluster status is ready before deploying nodes', 'Check node health before starting training').
Remove response JSON examples that don't add instructional value — Claude can infer response shapes from the tool definitions.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | Extremely verbose at ~500+ lines. Massive amounts of repetition — architecture patterns are shown in full examples AND repeated in the 'Architecture Patterns' section. Common use cases duplicate earlier examples. Response JSON examples add bulk without teaching Claude anything new. The skill could be reduced to 1/3 its size. | 1 / 3 |
Actionability | Every capability includes concrete, executable MCP tool calls with full parameter objects and expected response JSON. The examples are copy-paste ready with realistic configurations and cover the full lifecycle from training to inference to cluster management. | 3 / 3 |
Workflow Clarity | The distributed training section shows a logical sequence (init cluster → deploy nodes → connect → train → monitor → terminate), but there are no explicit validation checkpoints or error-recovery feedback loops. The 'Common Use Cases' section shows sequential steps but lacks 'validate before proceeding' gates. Troubleshooting is listed but not integrated into workflows. | 2 / 3 |
Progressive Disclosure | This is a monolithic wall of content with everything inline. The Architecture Patterns section, all the use case examples, and the extensive API reference content should be split into separate referenced files. The 'Related Skills' and 'Resources' sections hint at external references but the body itself is far too long for a SKILL.md overview. | 1 / 3 |
Total | 7 / 12 Passed |
Validation
81%Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.
Validation — 9 / 11 Passed
Validation for skill structure
| Criteria | Description | Result |
|---|---|---|
skill_md_line_count | SKILL.md is long (739 lines); consider splitting into references/ and linking | Warning |
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 9 / 11 Passed | |
- Repository
- majiayu000/claude-skill-registry-data
- Commit
3dd3ac0
- Reviewed
Table of Contents
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.