langchain

Framework for building LLM-powered applications with agents, chains, and RAG. Supports multiple providers (OpenAI, Anthropic, Google), 500+ integrations, ReAct agents, tool calling, memory management, and vector store retrieval. Use for building chatbots, question-answering systems, autonomous agents, or RAG applications. Best for rapid prototyping and production deployments.

Quality

73%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./backend/cli/skills/llm-tools/langchain/SKILL.md

Quality

Content

63%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 runnable code for every major feature area and a coherent learning progression. Its weaknesses are duplication — the main file re-inlines content the three reference files already provide — marketing-style padding, and a mix of deprecated and current APIs that could mislead on execution details.

Suggestions

Move the Vector stores, Document loaders, Text splitters, and LangSmith observability sections out of SKILL.md and into the existing references/integration.md and references/rag.md, leaving one-line pointers in the References section — this also removes the duplication dragging conciseness.

Cut the Metrics block, Performance benchmarks table, and LangChain vs LangGraph comparison (or compress them to two lines under 'When to use'), and strip hard-coded model version strings in favor of a note to check the current model ID.

Standardize on one API generation: replace the deprecated LLMChain / ConversationChain / RetrievalQA / load_qa_with_sources_chain examples with their current equivalents (or move them to an explicit 'legacy API' section), and fix the create_tool_calling_agent prompt argument to use a ChatPromptTemplate.

Dimension	Reasoning	Score
Conciseness	The body is mostly dense code with little concept over-explanation, but it carries padding Claude does not need ('119,000+ GitHub stars', '272,000+ repositories', a latency benchmark table, a LangChain-vs-LangGraph marketing comparison) plus time-sensitive model strings and version pins ('claude-sonnet-4-5-20250929', 'Version: 0.3+') that will age, fitting 'mostly efficient but includes some unnecessary explanation or could be tightened'.	3 / 5
Actionability	Nearly every section ships copy-paste-ready executable Python, but there are more than trivial gaps: 'prompt="Answer questions using available tools"' passes a string where create_tool_calling_agent requires a ChatPromptTemplate, 'loader.load("https://...")' misuses WebBaseLoader, and deprecated legacy APIs (LLMChain, ConversationChain, RetrievalQA, load_qa_with_sources_chain) are mixed with the current create_agent API, matching 'mostly executable guidance with minor gaps' rather than fully executable.	4 / 5
Workflow Clarity	The flow is logical and well-sequenced (when-to-use → install → quick start → core concepts → a RAG pipeline with numbered steps 1–6 → advanced patterns → best practices) and no risky batch/destructive operations require validation, though the mixed old/new API guidance leaves a minor gap in checkpoint clarity.	4 / 5
Progressive Disclosure	Three real, clearly signaled reference files exist and are linked with descriptions in a dedicated References section, but the ~470-line body inlines substantial material those files already cover (Vector stores, Document loaders, Text splitters, LangSmith observability all duplicate integration.md/rag.md), which is more than the 'minor organization gaps' of anchor 4 and fits 'content that should be separate is inline'.	3 / 5
	Total	14 / 20 Passed

Description

83%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.

A strong description: it clearly states what the framework does and gives an explicit 'Use for...' clause with concrete, natural trigger scenarios. It could be sharpened by converting feature-list nouns into actions and adding a few synonyms, but it sits comfortably above the midpoint on every dimension.

Dimension	Reasoning	Score
Specificity	Lists several specific capabilities ('agents, chains, and RAG', 'tool calling, memory management, and vector store retrieval', 'building chatbots, question-answering systems'), but many are feature nouns rather than concrete actions and 'Best for rapid prototyping and production deployments' is a claim rather than a capability, so it falls just short of the comprehensive anchor 5.	4 / 5
Completeness	Explicitly answers both 'what' ('Framework for building LLM-powered applications with agents, chains, and RAG...') and 'when' ('Use for building chatbots, question-answering systems, autonomous agents, or RAG applications') with concrete trigger phrases, matching the anchor-5 example rather than the weaker 'when' of anchor 4.	5 / 5
Trigger Term Quality	Includes natural user phrases like 'chatbots', 'RAG applications', 'autonomous agents', and 'question-answering systems', but misses common synonyms such as 'retrieval-augmented generation' spelled out, 'LLM app', or 'GenAI', so a few natural terms are missing per anchor 4.	4 / 5
Distinctiveness Conflict Risk	The niche is mostly distinct (multi-provider agent/RAG framework with 500+ integrations), but it overlaps with closely related skills like LlamaIndex or Haystack for RAG and chatbot work, fitting anchor 4 ('minor overlap risk with closely related skills') rather than anchor 5's minimal-conflict bar.	4 / 5
	Total	17 / 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
metadata_version	'metadata.version' is missing	Warning
frontmatter_unknown_keys	Unknown frontmatter key(s) found; consider removing or moving to metadata	Warning

	Total	14 / 16 Passed

Repository: synthetic-sciences/openscience
Path: backend/cli/skills/llm-tools/langchain/SKILL.md
Commit: 4082a2e

Reviewed: 1 day 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.