Langchain Context

LangChain context navigator

Overview

This navigator covers the LangChain Python ecosystem in depth, plus a thin orientation pointer for the JS port:

• Primary: the langchain-ai/langchain Python monorepo — langchain-core, the new langchain package (v1; built from libs/langchain_v1/), the legacy langchain-classic package (built from libs/langchain/), in-tree partner integrations, and the auxiliary libs (text-splitters, model-profiles, standard-tests).
• Python companion repos: LangGraph (langchain-ai/langgraph), LangSmith SDK (langchain-ai/langsmith-sdk), Deep Agents (langchain-ai/deepagents), and the out-of-tree partner monorepos langchain-ai/langchain-google and langchain-ai/langchain-aws. Plus the docs source at langchain-ai/docs.
• JS port (thin coverage only): langchain-ai/langchainjs is included as a pointer reference for orienting JS questions — full JS depth is out of scope; follow source links into the JS repo for non-trivial questions.

When asked a question this navigator's domain covers:

1. Scan the Catalog below for the matching topic.
2. Follow the link to read the reference file.
3. If the question spans multiple references, consult the Cross-reference map.
4. If a reference points at a source URL for deeper detail, follow it only if the reference itself didn't answer the question.

For first-time orientation — what packages exist, how langchain vs langchain-classic differ, where companion repos live — start with langchain-overview.md.

Claims policy

Cite by default, and make load-bearing claims verifiable:

1. Inline-cite every load-bearing claim with its SHA-pinned permalink — the https://github.com/langchain-ai/langchain/blob/<sha>/libs/langchain_v1/langchain/agents/factory.py#L17-L25-style link the reference gives for that fact (versions, defaults, signatures, deprecations, behavior a user could get wrong by guessing). Put the permalink inline, on the claim. Use a bare filename parenthetical (e.g. (langchain-v1-agents.md)) only when the reference genuinely provides no permalink. This inline permalink is what the grounded-citation eval (SELF-AUDIT Check 8) grades.
2. Don't cite orientational prose — "what is X?", "when did X launch?" — answer those from this navigator alone; opening a reference is itself a citation gesture.
3. End with a one-line provenance footer, emitted italic, formatted *References consulted: foo.md, bar.md. Grounded in {{LIBRARY}}@{{VERSION}} — [reference index]({{INDEX_URL}}).* The footer is a summary of what you read — not a substitute for the inline permalinks on the claims. The {{LIBRARY}} / {{VERSION}} / {{INDEX_URL}} tokens are agent-substituted at answer time, so they appear literally in the stamped SKILL.md.
4. If no reference was opened, say so in the footer ("Answered from general knowledge — no {{LIBRARY}} references consulted") — never fake it.

The voice is competent and careful — no "as an AI assistant" hedging.

Catalog

Main monorepo (langchain-ai/langchain)

Python companion repos

JS port (thin pointer only)

Cross-reference map

• Building a RAG pipeline → start at langchain-core-retrieval for the protocol shape; cross to langchain-text-splitters for the chunking step and langchain-core-runnables for composing the prompt-model-parser tail with LCEL.
• Building an agent → start at langchain-v1-agents for create_agent and middleware; cross to langchain-core-models for tool definitions and structured output, to langgraph-overview for the underlying graph runtime (create_agent returns a StateGraph), and to langchain-classic only if the user has 0.x AgentExecutor code to migrate.
• Building a "batteries-included" agent (planning, filesystem, sub-agents) → start at deepagents-overview for create_deep_agent; cross to langchain-v1-agents for the underlying create_agent it composes with, and to langgraph-overview for streaming/persistence.
• Durable agents / pause-resume / human-in-the-loop / checkpoints → start at langgraph-overview for the LangGraph checkpoint and interrupt surface (the langchain repo doesn't implement these — LangGraph does). For AWS-managed checkpoint backends specifically, cross to langchain-aws-overview (langgraph-checkpoint-aws).
• Adding support for a new model provider → start at langchain-partners for the in-tree package layout; cross to langchain-core-models for the BaseChatModel contract details. For Google or AWS specifically, those are out-of-tree — see langchain-google-overview or langchain-aws-overview.
• Observability / tracing / streaming UI → start at langchain-core-callbacks for the event surface; cross to langchain-core-runnables for how astream_events is wired into the Runnable protocol. For LangSmith specifically — when you'd reach for the SDK directly vs. let env vars do the work — cross to langsmith-sdk-overview.
• Migrating 0.x code → start at langchain-classic for what moved where; cross to langchain-v1-agents for the v1 replacement of initialize_agent + AgentExecutor + memory, and to langgraph-overview if the user is asking about graph-level customization the v1 agent surface doesn't expose.
• "Why is LLMChain deprecated / what replaces it" → start at langchain-classic; cross to langchain-core-runnables for the LCEL replacement pattern.
• "Where is the docs page for X / where do I file a docs issue" → start at docs-overview for the docs.langchain.com source layout, the packages.yml registry, and the distinction between docs.langchain.com vs. reference.langchain.com.
• "What's the JS equivalent of " → start at langchainjs-overview for the npm package mapping; for non-trivial JS questions follow the source links into the JS repo and JS docs.

Instructions to Claude

When loading a reference file, the path syntax depends on the platform:

• Claude Code: Read the reference using the platform-provided skill-directory variable: Read $CLAUDE_SKILL_DIR/references/<source-slug>-<topic>.md

• Claude Desktop: Read the reference using a relative path; the platform resolves it from the skill's installed location: Read references/<source-slug>-<topic>.md

Loading rules:

• Load one reference at a time unless the Cross-reference map says to load both.
• If the primary reference doesn't fully answer the question, follow any source URL pointers it provides for deeper detail.
• Do not eagerly load companion files; only follow companion links when the primary reference says to.
• If the user's question is clearly out of scope for this contextualizer, don't invoke this skill at all.
• The repo's two same-looking package directories (libs/langchain/ and libs/langchain_v1/) publish to PyPI under different names (langchain-classic and langchain respectively). Always disambiguate when discussing them — langchain-overview and the first sections of langchain-v1-agents and langchain-classic cover this.

Progressive disclosure

References prioritize curated insight over re-specifying upstream sources:

• Gotchas, cross-system patterns, and "why" context are kept in the reference (curation value).
• Exact schemas, API signatures, and parameter lists are summarized in the reference and linked to their authoritative source via SHA-pinned URLs.

When a reference includes a source URL pointer, follow it only when the reference's own summary didn't cover the question. The contextualizer is optimized for the common case; the upstream source is the long tail.

File URLs in references are pinned to per-source SHAs at the time of the last /skill-engine:refresh run (2026-05-18). Each reference's links point at the SHA captured when that source was discovered. Run /skill-engine:refresh to re-pin against current HEAD across all sources. Per-source SHAs are recorded in research/source-paths.json.

Source: nick-railsback/skill-engine — distributed by TomeVault.