Name: llm-app-patterns
Rating: 47.2 (1 reviews)
Author: administrakt0r

llm-app-patterns

Production-ready patterns for building LLM applications, inspired by [Dify](https://github.com/langgenius/dify) and industry best practices.

Quality

51%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./plugins/AI-Agents-Safe-Coding-Skills-claude/skills/llm-app-patterns/SKILL.md

The canonical home for this skill is llm-app-patterns in administrakt0r/AI-Agents-Safe-Coding-Skills

Quality

Content

57%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 a rich, actionable catalog of LLM application patterns with strong code examples and a useful decision matrix. Its main weaknesses are that all detailed reference material is inlined into SKILL.md with no progressive disclosure, and there are no validation/feedback checkpoints framing the patterns as guided workflows.

Suggestions

Move the bulk reference material (vector DB/embedding model option tables, full agent class implementations, metrics dictionaries) into reference files under references/ and keep SKILL.md as an overview with one-level-deep links.

Fix the broken line `output = step"parser"` in PromptChain.run and either implement or explicitly mark the placeholder helpers (llm, vector_db, bm25_search) as assumed interfaces.

Add brief validation/feedback-loop notes for the operations most prone to silent failure (e.g. verify retrieved context before generation, validate prompt template variables, confirm cache determinism) to lift workflow clarity.

Dimension	Reasoning	Score
Conciseness	The body is code-dense with little prose padding, but ~700 lines of comprehensive reference material (vector DB options, embedding models, full class implementations) are inlined in SKILL.md where tighter overview-plus-reference separation would reduce token load.	3 / 5
Actionability	Extensive concrete, mostly-executable Python covers RAG, agents, prompt templating, caching, retry, and fallback; minor gaps include the broken `step"parser"` line and reliance on undefined helpers like `llm` and `vector_db`.	4 / 5
Workflow Clarity	This is a patterns catalog with a helpful decision matrix rather than a sequenced workflow, and it lacks explicit validation checkpoints or feedback loops for the operations it describes.	3 / 5
Progressive Disclosure	Section headers give good structure, but no bundle/reference files exist and all detailed pattern implementations are inlined in SKILL.md rather than split into one-level-deep references.	3 / 5
	Total	13 / 20 Passed

Description

45%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 domain and purpose but stays generic: it lists no concrete capabilities and provides no 'when to use' trigger guidance. It reads more like a tagline than an activatable skill description.

Suggestions

Replace the generic 'patterns for building LLM applications' with concrete capabilities, e.g. 'Design RAG pipelines, build ReAct/function-calling agents, version prompts, and instrument LLMOps'.

Add an explicit trigger clause, e.g. 'Use when designing LLM-powered apps, implementing RAG, building tool-using agents, or setting up LLM monitoring.'

Surface the natural trigger terms users say (RAG, agents, LLMOps, prompt chaining) so the skill is distinguishable from neighbouring LLM skills.

Dimension	Reasoning	Score
Specificity	The description names the domain ('building LLM applications') but offers only one generic action and no concrete capabilities like RAG, agents, or LLMOps, matching the 'names the domain but actions are minimal' anchor.	2 / 5
Completeness	It gives a clear 'what' ('Production-ready patterns for building LLM applications') but includes no 'Use when...' trigger clause, capping completeness at 3 per the rubric guidance.	3 / 5
Trigger Term Quality	'LLM applications' is a natural user term and 'Dify' a recognizable platform name, but common variations users actually say (RAG, agents, LLMOps) are absent.	3 / 5
Distinctiveness Conflict Risk	'Building LLM applications' is a broad umbrella that overlaps RAG, agent, prompt, and LLMOps skills, though it is a recognizable niche rather than fully generic.	3 / 5
	Total	11 / 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 (764 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: administrakt0r/AI-Agents-Safe-Coding-Skills
Path: plugins/AI-Agents-Safe-Coding-Skills-claude/skills/llm-app-patterns/SKILL.md
Commit: 57c135b

Reviewed: 4 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.