ai-ml

AI and machine learning workflow covering LLM application development, RAG implementation, agent architecture, ML pipelines, and AI-powered features.

Quality

26%

Does it follow best practices?

Impact

—

No eval scenarios have been run

Securityby

Passed

No known issues

Optimize this skill with Tessl

npx tessl skill review --optimize ./skills/antigravity-ai-ml/SKILL.md

Quality

Content

20%

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

This skill is a high-level orchestration document that reads more like a table of contents than actionable guidance. It lists dozens of sub-skills and generic action items without providing any concrete implementation details, code examples, or specific technical guidance. The repetitive phase structure inflates token usage significantly while delivering minimal value beyond what Claude could infer from the skill names alone.

Suggestions

Replace vague action items (e.g., 'Choose appropriate models', 'Configure retrieval') with concrete decision criteria, specific commands, or executable code examples that add real value.

Add validation checkpoints and feedback loops between phases—e.g., 'Verify RAG retrieval accuracy exceeds X% before proceeding to agent development'.

Dramatically reduce verbosity by consolidating the seven identically-structured phases into a concise reference table mapping phases to skills, with detailed guidance only where Claude needs non-obvious information.

Include at least one concrete, end-to-end example (e.g., a minimal RAG pipeline setup) with executable code to demonstrate the workflow in practice.

Dimension	Reasoning	Score
Conciseness	Extremely verbose and repetitive. The skill is essentially a long list of sub-skill references, vague action items, and checklists that provide no information Claude doesn't already know. Phrases like 'Define AI use cases', 'Choose appropriate models', and 'Design system architecture' are generic platitudes. The 'Copy-Paste Prompts' sections are trivial one-liners repeated across every phase. The entire document could be condensed to a fraction of its size.	1 / 3
Actionability	No concrete code, commands, or executable guidance anywhere. Every 'Actions' section is a list of vague directives like 'Set up API access', 'Implement chunking strategy', 'Configure retrieval'. The copy-paste prompts are just 'Use @skill-name to do X' which provide no real implementation detail. There are no examples, no code snippets, no specific configurations.	1 / 3
Workflow Clarity	The phases are sequenced logically (design → integration → RAG → agents → ML → observability → security), which provides some structural clarity. However, there are no validation checkpoints, no feedback loops, no error recovery steps, and no criteria for when to move between phases. The checklists at the end are generic and disconnected from the workflow steps.	2 / 3
Progressive Disclosure	The skill references many sub-skills by name, providing a one-level-deep structure. However, no bundle files are provided, so the references cannot be verified. The main document itself is a wall of repetitive content that could benefit from better organization—the seven phases follow an identical template that becomes tedious. The references are present but not clearly signaled with links or paths.	2 / 3
	Total	6 / 12 Passed

Description

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 broad AI/ML domain and lists several sub-areas but reads more like a category label than an actionable skill description. It lacks concrete actions (verbs), explicit trigger guidance ('Use when...'), and natural user-facing language variations. The overly broad scope also creates conflict risk with more specialized skills.

Suggestions

Add an explicit 'Use when...' clause specifying trigger scenarios, e.g., 'Use when the user asks about building LLM applications, implementing RAG pipelines, designing agent architectures, or adding AI-powered features to their project.'

Replace category labels with concrete actions using third-person verbs, e.g., 'Scaffolds LLM application architectures, implements retrieval-augmented generation with vector databases, designs multi-agent workflows, and builds ML training pipelines.'

Include common natural language trigger terms users would actually say, such as 'chatbot', 'embeddings', 'vector database', 'prompt engineering', 'model serving', 'fine-tuning', or 'OpenAI/Anthropic API integration'.

Dimension	Reasoning	Score
Specificity	Names the domain (AI/ML) and lists several areas like 'LLM application development, RAG implementation, agent architecture, ML pipelines,' but these are high-level categories rather than concrete actions. No specific verbs describing what the skill actually does (e.g., 'builds', 'configures', 'deploys').	2 / 3
Completeness	Describes 'what' at a high level (AI/ML workflow covering several areas) but completely lacks any 'when' clause or explicit trigger guidance. Per the rubric, a missing 'Use when...' clause caps completeness at 2, and the 'what' is also quite vague, warranting a score of 1.	1 / 3
Trigger Term Quality	Includes relevant keywords like 'LLM', 'RAG', 'agent architecture', 'ML pipelines', and 'AI-powered features' that users might mention. However, it misses common variations and natural phrases users would say, such as 'chatbot', 'embeddings', 'vector database', 'fine-tuning', 'prompt engineering', 'model training', or 'inference'.	2 / 3
Distinctiveness Conflict Risk	The scope is extremely broad—covering LLM development, RAG, agents, ML pipelines, and AI features—which could easily overlap with more specialized skills for any of those individual areas. The breadth makes it somewhat distinctive as a catch-all AI skill but increases conflict risk with narrower skills.	2 / 3
	Total	7 / 12 Passed

Validation

90%

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 — 10 / 11 Passed

Validation for skill structure

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

	Total	10 / 11 Passed

Repository: boisenoise/skills-collections
Commit: aff357c

Reviewed: 5 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.