agent-neural-network

Agent skill for neural-network - invoke with $agent-neural-network

1.60x

Quality

Does it follow best practices?

Impact

96%

1.60x

Average score across 3 eval scenarios

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./.agents/skills/agent-neural-network/SKILL.md

Quality

Discovery

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

This is an extremely weak description that provides virtually no useful information for skill selection. It lacks any concrete actions, trigger terms, use-case guidance, or distinguishing details. It reads more like a label than a description and would be nearly impossible for Claude to correctly select from a pool of skills.

Suggestions

Add specific concrete actions the skill performs, e.g., 'Builds, trains, and evaluates neural network architectures including CNNs, RNNs, and transformers.'

Add an explicit 'Use when...' clause with natural trigger terms, e.g., 'Use when the user asks about building neural networks, training deep learning models, configuring layers, or optimizing model performance.'

Remove the invocation syntax from the description (it's metadata, not a capability description) and replace with domain-specific keywords users would naturally use, such as 'deep learning', 'model training', 'backpropagation', 'PyTorch', 'TensorFlow'.

Dimension	Reasoning	Score
Specificity	The description contains no concrete actions whatsoever. It only says 'Agent skill for neural-network' which is entirely vague and abstract, providing no information about what the skill actually does.	1 / 3
Completeness	The description fails to answer both 'what does this do' and 'when should Claude use it'. There is no explanation of capabilities and no 'Use when...' clause or equivalent trigger guidance.	1 / 3
Trigger Term Quality	The only keyword present is 'neural-network', which is a broad technical term. There are no natural user-facing trigger terms like 'train model', 'classify', 'deep learning', 'layers', etc. The invocation syntax '$agent-neural-network' is not a natural user phrase.	1 / 3
Distinctiveness Conflict Risk	The description is so generic that 'neural-network' could overlap with any ML, AI, deep learning, or data science skill. There are no distinct triggers or specific use cases to differentiate it.	1 / 3
	Total	4 / 12 Passed

Implementation

14%

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

This skill reads more like a persona description or role-play prompt than an actionable skill document. It spends most of its tokens describing what the agent is and what it knows (neural network types, quality standards, advanced capabilities) rather than providing concrete, executable guidance for specific tasks. The three MCP tool call examples are the only actionable content, but they lack context on when to use each, expected outputs, and error handling.

Suggestions

Remove all descriptive content about neural network types and ML concepts that Claude already knows, and replace with concrete task-specific workflows (e.g., 'To train a classification model: step 1, step 2...')

Add explicit validation checkpoints to workflows, such as checking training metrics after each epoch, validating model outputs before deployment, and handling common failure modes

Add expected output examples for each MCP tool call so Claude knows what a successful response looks like and how to handle errors

Restructure as a concise quick-reference with tool signatures and common patterns, moving any detailed architecture guides to separate referenced files

Dimension	Reasoning	Score
Conciseness	Extremely verbose with extensive explanations of concepts Claude already knows (neural network architectures, ML workflows, quality standards). Lists of architecture types, quality standards, and advanced capabilities are padded descriptions that don't add actionable value. The bulk of the content is descriptive rather than instructive.	1 / 3
Actionability	The JavaScript code examples showing MCP tool calls are concrete and provide specific parameter structures, which is useful. However, the rest of the skill is vague guidance ('always consider scalability, reproducibility') with no executable workflows, no specific commands for common tasks, and no concrete examples of expected outputs or error handling.	2 / 3
Workflow Clarity	The 6-step ML workflow is a high-level abstract description with no concrete commands, validation checkpoints, or feedback loops. There's no guidance on what to do when training fails, how to validate models, or how to recover from errors. For operations involving distributed training and deployment, this lack of validation steps is a significant gap.	1 / 3
Progressive Disclosure	The content is a monolithic wall of text with no references to external files, no clear navigation structure, and no separation of overview from detailed content. Everything is dumped into a single file with no bundle files to support it. Lists of architectures and capabilities that could be referenced separately are all inline.	1 / 3
	Total	5 / 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 — 11 / 11 Passed

Validation for skill structure

No warnings or errors.

Repository: ruvnet/claude-flow
Commit: ca77f83

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.