langchain-sdk-patterns
Apply production-ready LangChain SDK patterns for structured output, fallbacks, batch processing, streaming, and caching. Trigger: "langchain SDK patterns", "langchain best practices", "idiomatic langchain", "langchain architecture", "withStructuredOutput", "withFallbacks", "abatch".
67
82%
Does it follow best practices?
Impact
—
No eval scenarios have been run
Passed
No known issues
Quality
Discovery
100%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 a strong skill description that clearly identifies its domain (LangChain SDK), lists specific capabilities (structured output, fallbacks, batch processing, streaming, caching), and provides explicit trigger terms covering both natural language and API-specific keywords. The description is concise, uses third-person voice, and would be easily distinguishable from other skills in a large skill library.
| Dimension | Reasoning | Score |
|---|---|---|
Specificity | Lists multiple specific concrete actions: structured output, fallbacks, batch processing, streaming, and caching. These are clearly defined capabilities within the LangChain SDK domain. | 3 / 3 |
Completeness | Clearly answers 'what' (apply production-ready LangChain SDK patterns for structured output, fallbacks, batch processing, streaming, and caching) and 'when' via an explicit 'Trigger:' clause listing specific trigger terms. The trigger clause serves the same function as a 'Use when...' clause. | 3 / 3 |
Trigger Term Quality | Includes a strong mix of natural language terms ('langchain best practices', 'langchain architecture') and specific technical terms users would actually type ('withStructuredOutput', 'withFallbacks', 'abatch', 'idiomatic langchain'). Good coverage of how users would phrase requests. | 3 / 3 |
Distinctiveness Conflict Risk | Highly distinctive — focuses specifically on LangChain SDK patterns with named methods like 'withStructuredOutput', 'withFallbacks', and 'abatch'. Unlikely to conflict with general coding skills or other framework-specific skills. | 3 / 3 |
Total | 12 / 12 Passed |
Implementation
64%Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.
This is a solid reference-style skill with highly actionable, executable code examples covering important LangChain patterns. Its main weaknesses are that it's too long for a single SKILL.md (would benefit from splitting patterns into separate files), lacks integrated validation/error-recovery workflows within the code examples, and includes some content Claude could generate independently (like the manual retry wrapper). The mixed TypeScript/Python without explanation is slightly jarring.
Suggestions
Split the 7 patterns into separate referenced files (e.g., patterns/structured-output.md) and keep SKILL.md as a concise overview with one-line descriptions and links to each pattern.
Add inline error handling and validation steps within the code examples — e.g., show try/catch around structured output with ZodError recovery, or partial failure handling in batch processing.
Remove or consolidate the manual retry wrapper (Pattern 5) since the built-in maxRetries config is sufficient and Claude can generate custom retry logic on demand.
Either commit to TypeScript throughout or explicitly note the Python caching pattern as a Python-only feature with a brief justification for the language switch.
| Dimension | Reasoning | Score |
|---|---|---|
Conciseness | The content is mostly efficient with executable code examples, but includes some unnecessary elements — Pattern 6 switches to Python without justification, the anti-patterns and error handling tables add moderate value but could be tighter, and some patterns (like the manual retry wrapper) are things Claude could generate on demand. The overall length (~150 lines of code examples across 7 patterns) is borderline verbose for a skill file. | 2 / 3 |
Actionability | Every pattern provides fully executable, copy-paste-ready code with correct imports, typed schemas, and realistic examples. The code is concrete and specific — not pseudocode — with inline comments explaining key behaviors (e.g., cache hit timing, concurrency control). | 3 / 3 |
Workflow Clarity | Each pattern is presented as an independent recipe rather than a sequenced workflow. There are no validation checkpoints or feedback loops — for example, the structured output pattern doesn't show what to do when ZodError occurs inline, and the batch processing pattern lacks error handling for partial failures. The error handling table partially compensates but is disconnected from the code examples. | 2 / 3 |
Progressive Disclosure | The content is a monolithic file with 7 patterns, 2 tables, and resource links all inline. With no bundle files, the patterns could benefit from being split into separate reference files with the SKILL.md serving as a concise overview. The external links to LangChain docs and the 'Next Steps' reference to another skill are good, but the body itself is too long for a single skill file. | 2 / 3 |
Total | 9 / 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 |
|---|---|---|
allowed_tools_field | 'allowed-tools' contains unusual tool name(s) | Warning |
frontmatter_unknown_keys | Unknown frontmatter key(s) found; consider removing or moving to metadata | Warning |
Total | 9 / 11 Passed | |
- Commit
a04d1a2
- 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.