n8n-code-tool

Name: n8n-code-tool
Author: czlonkowski

199installs983forks

How to install n8n-code-tool?

czlonkowski/n8n-skills/n8n-code-tool

$npx -y skills add czlonkowski/n8n-skills --skill n8n-code-tool

Installs into the current project.

Summary

Write JavaScript or Python for the n8n Custom Code Tool (@n8n/n8n-nodes-langchain.toolCode) — the AI-agent-callable tool, NOT the workflow Code node. Use when building a Code Tool attached to an AI Agent, writing code that an LLM will invoke, parsing the query input, returning

Files · 1

View on GitHub

SKILL.md

1	# n8n Custom Code Tool
2
3	Expert guidance for writing code inside `@n8n/n8n-nodes-langchain.toolCode` — the tool an AI Agent can invoke, not the regular workflow Code node.
4
5	---
6
7	## ⚠️ This is NOT the Code node
8
9	The Custom Code Tool looks like a Code node in the editor — same JavaScript editor, similar layout — but it is a completely different node from a different package with a different runtime contract.
10
11	\| \| Code node \| Custom Code Tool \|
12	\|---\|---\|---\|
13	\| Node type \| `n8n-nodes-base.code` \| `@n8n/n8n-nodes-langchain.toolCode` \|
14	\| Package \| `n8n-nodes-base` \| `@n8n/n8n-nodes-langchain` \|
15	\| Invoked by \| Previous node (workflow flow) \| AI Agent (LangChain) \|
16	\| Input \| `$input.all()` — item stream \| `query` — string or object from LLM \|
17	\| Return \| `[{json: {...}}]` (items array) \| A string \|
18	\| `$fromAI()` \| N/A \| Not available (see Errors) \|
19	\| HTTP helper \| `this.helpers.httpRequest` (auth helpers blocked) \| Not exposed to the tool sandbox \|
20	\| State \| Per-run execution data \| No `getContext`, no `$getWorkflowStaticData` \|
21
22	If you treat it like a Code node, it fails. The rest of this skill covers the Code Tool's actual contract.
23
24	---
25
26	## Quick Start
27
28	### Minimal JavaScript Code Tool
29
30	```javascript
31	// `query` is whatever the AI sent (a string by default)
32	return `You asked: ${query}`;
33	```
34
35	### Minimal Python Code Tool
36
37	```python
38	# `_query` is whatever the AI sent (a string by default)
39	return f"You asked: {_query}"
40	```
41
42	### Essential Rules
43
44	1. Return a string. Numbers are auto-converted. Anything else throws `"The response property should be a string, but it is an object"`.
45	2. Input variable is fixed: `query` (JS), `_query` (Python). You cannot rename it.
46	3. Do NOT use `$fromAI()` inside the Code Tool sandbox — it throws `"No execution data available"`.
47	4. Do NOT use `[{json: {...}}]` return format — that's for Code nodes. Throws `"Wrong output type returned"`.
48	5. Use a descriptive tool name (letters/numbers/underscores, v1.1+). The agent calls the tool by its name.
49	6. Write a precise description — the LLM decides whether to invoke the tool based on it.
50
51	---
52
53	## The Two Input Modes
54
55	The Code Tool has two input shapes, controlled by `specifyInputSchema`:
56
57	### Mode 1: Unstructured (default, `specifyInputSchema: false`)
58
59	The AI passes a single string as `query`. If you need multiple fields, the AI has to stuff them into that one string and you parse them out. In practice, LLMs will happily pass a JSON string if your description tells them to.
60
61	```javascript
62	// Parse a JSON string the AI sent
63	let params;
64	try {
65	params = typeof query === 'string' ? JSON.parse(query) : query;
66	} catch (e) {
67	throw new Error('Expected a JSON object. Parser said: ' + e.message);
68	}
69	const price = Number(params.price);
70	const months = Number(params.months);
71	// ...
72	return JSON.stringify({ monthly_payment: /* ... */ });
73	```
74
75	Pros: simplest to set up, one field to describe.
76	Cons: no schema validation — if the LLM forgets a field, the tool throws at runtime.
77
78	Best for: quick prototypes, tools with one natural input (a question, a URL, a text blob).
79
80	### Mode 2: Structured (`specifyInputSchema: true`)
81
82	The tool becomes a LangChain `DynamicStructuredTool`. The LLM sees a typed argument schema and passes a validated object as `query`. You access fields directly.
83
84	```javascript
85	// query is now an object matching your schema
86	const price = query.price;
87	const months = query.months;
88	const residual_percent = query.residual_percent;
89
90	const monthly = computeAnnuity(price, months, residual_percent);
91	return JSON.stringify({ monthly_payment: monthly });
92	```
93
94	Schema is defined via either:
95	- `schemaType: "fromJson"` + `jsonSchemaExample` (n8n v≥1.3) — paste an example JSON, n8n infers the schema
96	- `schemaType: "manual"` + `inputSchema` — write a full JSON Schema yourself
97
98	Pros: LLM gets type hints, invalid calls rejected before your code runs, cleaner code.
99	Cons: a little more setup; requi

People who liked this also viewed

6 picks

Type

n8n-code-tool

199installs983forks

View on GitHub

How to install n8n-code-tool?

czlonkowski/n8n-skills/n8n-code-tool

$npx -y skills add czlonkowski/n8n-skills --skill n8n-code-tool

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# n8n Custom Code Tool
2
3	Expert guidance for writing code inside `@n8n/n8n-nodes-langchain.toolCode` — the tool an AI Agent can invoke, not the regular workflow Code node.
4
5	---
6
7	## ⚠️ This is NOT the Code node
8
9	The Custom Code Tool looks like a Code node in the editor — same JavaScript editor, similar layout — but it is a completely different node from a different package with a different runtime contract.
10
11	\| \| Code node \| Custom Code Tool \|
12	\|---\|---\|---\|
13	\| Node type \| `n8n-nodes-base.code` \| `@n8n/n8n-nodes-langchain.toolCode` \|
14	\| Package \| `n8n-nodes-base` \| `@n8n/n8n-nodes-langchain` \|
15	\| Invoked by \| Previous node (workflow flow) \| AI Agent (LangChain) \|
16	\| Input \| `$input.all()` — item stream \| `query` — string or object from LLM \|
17	\| Return \| `[{json: {...}}]` (items array) \| A string \|
18	\| `$fromAI()` \| N/A \| Not available (see Errors) \|
19	\| HTTP helper \| `this.helpers.httpRequest` (auth helpers blocked) \| Not exposed to the tool sandbox \|
20	\| State \| Per-run execution data \| No `getContext`, no `$getWorkflowStaticData` \|
21
22	If you treat it like a Code node, it fails. The rest of this skill covers the Code Tool's actual contract.
23
24	---
25
26	## Quick Start
27
28	### Minimal JavaScript Code Tool
29
30	```javascript
31	// `query` is whatever the AI sent (a string by default)
32	return `You asked: ${query}`;
33	```
34
35	### Minimal Python Code Tool
36
37	```python
38	# `_query` is whatever the AI sent (a string by default)
39	return f"You asked: {_query}"
40	```
41
42	### Essential Rules
43
44	1. Return a string. Numbers are auto-converted. Anything else throws `"The response property should be a string, but it is an object"`.
45	2. Input variable is fixed: `query` (JS), `_query` (Python). You cannot rename it.
46	3. Do NOT use `$fromAI()` inside the Code Tool sandbox — it throws `"No execution data available"`.
47	4. Do NOT use `[{json: {...}}]` return format — that's for Code nodes. Throws `"Wrong output type returned"`.
48	5. Use a descriptive tool name (letters/numbers/underscores, v1.1+). The agent calls the tool by its name.
49	6. Write a precise description — the LLM decides whether to invoke the tool based on it.
50
51	---
52
53	## The Two Input Modes
54
55	The Code Tool has two input shapes, controlled by `specifyInputSchema`:
56
57	### Mode 1: Unstructured (default, `specifyInputSchema: false`)
58
59	The AI passes a single string as `query`. If you need multiple fields, the AI has to stuff them into that one string and you parse them out. In practice, LLMs will happily pass a JSON string if your description tells them to.
60
61	```javascript
62	// Parse a JSON string the AI sent
63	let params;
64	try {
65	params = typeof query === 'string' ? JSON.parse(query) : query;
66	} catch (e) {
67	throw new Error('Expected a JSON object. Parser said: ' + e.message);
68	}
69	const price = Number(params.price);
70	const months = Number(params.months);
71	// ...
72	return JSON.stringify({ monthly_payment: /* ... */ });
73	```
74
75	Pros: simplest to set up, one field to describe.
76	Cons: no schema validation — if the LLM forgets a field, the tool throws at runtime.
77
78	Best for: quick prototypes, tools with one natural input (a question, a URL, a text blob).
79
80	### Mode 2: Structured (`specifyInputSchema: true`)
81
82	The tool becomes a LangChain `DynamicStructuredTool`. The LLM sees a typed argument schema and passes a validated object as `query`. You access fields directly.
83
84	```javascript
85	// query is now an object matching your schema
86	const price = query.price;
87	const months = query.months;
88	const residual_percent = query.residual_percent;
89
90	const monthly = computeAnnuity(price, months, residual_percent);
91	return JSON.stringify({ monthly_payment: monthly });
92	```
93
94	Schema is defined via either:
95	- `schemaType: "fromJson"` + `jsonSchemaExample` (n8n v≥1.3) — paste an example JSON, n8n infers the schema
96	- `schemaType: "manual"` + `inputSchema` — write a full JSON Schema yourself
97
98	Pros: LLM gets type hints, invalid calls rejected before your code runs, cleaner code.
99	Cons: a little more setup; requi

People who liked this also viewed

6 picks

Type