Atlas / MCP servers / wasintoh / LINE OA

LINE OACAUTION

Name: LINE OA
Author: wasintoh

mcp/wasintoh/line-oa

MCP server by wasintoh · ai agent · claude · claude mcp · Audited 2026-10-07

Talk to your LINE Official Account through AI — broadcasts, Flex & tappable Rich Messages, Thai-ready rich menus, coupons, insights, LINE Shopping. New in 2.2: attach a picture, say send — zero hosting setup, verified before success. 51 tools (37+14), multi-OA. Works with Claude Cowork, Claude Code,

Verdict

CAUTION

Grade

Trust score

86 /100

Exposed tools

51 33r · 15w · 3d

Transport

stdio · streamable-http

License

MIT

Stars

100

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

MCP Ultimate 2.2.1 — MCP server for LINE Official Account" width="720">

Operate your LINE Official Account from any AI agent — through natural language.

🛠️ Fixed in 2.2.1: run the MCP on your own server and point any AI agent at one URL — it now holds up through real sessions and full broadcasts (no more instant "Internal server error" in HTTP mode), and hosted pictures now cache properly, so thousands of customers opening your broadcast at once barely touch your machine. Thanks to a production field report by Norapat Limpagan! 🙏

🖼️ New in 2.2: attach a picture, say "send this as a Rich Message" — and it appears in your customers' chats. No hosting. No image resizing. No setup.

[](https://github.com/wasintoh/line-oa-mcp-ultimate/actions/workflows/test.yml) [](https://www.npmjs.com/package/line-oa-mcp-ultimate) [](https://www.npmjs.com/package/line-oa-mcp-ultimate) [](.github/workflows/test.yml) [](https://www.npmjs.com/package/line-oa-mcp-ultimate) [](LICENSE) [](#requirements) [](https://modelcontextprotocol.io)

Quick Start · What you can do · Configuration · Security · [Documentation](#docume

Read from source at commit 90dcad810469OBSERVED · 2026-10-07

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code

claude mcp add line-oa-mcp-ultimate --env LINE_CHANNEL_ACCESS_TOKEN=${LINE_CHANNEL_ACCESS_TOKEN} --env LINE_CHANNEL_SECRET=${LINE_CHANNEL_SECRET} --env LINE_MYSHOP_API_KEY=${LINE_MYSHOP_API_KEY} --env MCP_HTTP_TOKEN=${MCP_HTTP_TOKEN} -- npx -y [email protected]

claude-desktop

{
  "mcpServers": {
    "line-oa-mcp-ultimate": {
      "command": "npx",
      "args": [
        "-y",
        "[email protected]"
      ],
      "env": {
        "LINE_CHANNEL_ACCESS_TOKEN": "${LINE_CHANNEL_ACCESS_TOKEN}",
        "LINE_CHANNEL_SECRET": "${LINE_CHANNEL_SECRET}",
        "LINE_MYSHOP_API_KEY": "${LINE_MYSHOP_API_KEY}",
        "MCP_HTTP_TOKEN": "${MCP_HTTP_TOKEN}"
      }
    }
  }
}

Exposed tools (51)

33 read · 15 write · 3 destructive. Blast radius: 3 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`line_audit_user_menu`	read
`line_build_audience_from_csv`	read
`line_build_audience_from_engagement`	read
`line_build_rich_menu`	read
`line_cancel_order`	write
`line_check_token`	read
`line_create_checkout_link`	write
`line_create_product`	write
`line_delete_audience`	destructive
`line_delete_product`	destructive
`line_delete_rich_menu`	destructive
`line_design_card`	read
`line_design_flex`	read
`line_design_imagemap`	read
`line_design_rich_menu_image`	read
`line_estimate_send_cost`	write
`line_find_sticker`	read
`line_fulfill_order`	write
`line_get_coupon_stats`	read
`line_get_message_stats`	read
`line_get_narrowcast_progress`	read
`line_get_oa_report`	read
`line_get_oa_status`	read
`line_get_order`	write
`line_get_settlement`	read
`line_get_user_profile`	read
`line_image_host_status`	read
`line_link_rich_menu`	read
`line_list_audiences`	read
`line_list_followers`	read
`line_list_oas`	read
`line_list_orders`	read
`line_list_products`	read
`line_list_rich_menus`	read
`line_manage_coupon`	read
`line_manage_inventory`	read
`line_manage_liff_app`	read
`line_manage_product_variant`	read
`line_manage_rich_menu_alias`	read
`line_prepare_image`	read
`line_run_on_many_oas`	write
`line_send_message`	write
`line_set_default_rich_menu`	write
`line_set_product_visibility`	write
`line_set_webhook_endpoint`	write
`line_test_webhook`	read
`line_update_audience`	write
`line_update_product`	write
`line_update_product_price`	write
`line_upload_rich_menu_image`	write
`line_use_oa`	read

Trust audit

CAUTIONgrade B · trust 86/100 Install with care. The audit found things worth knowing before you trust its output.

Layer	What it checks	Result
L0	Provenance & inventory	PASS
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (5 observation(s))
Network: declared (5 observation(s))
Shell: declared (1 observation(s))
Dependencies: not all pinned
Secrets in source: none-found

Findings (22)

HIGHNetwork egress · net.metadata_ip · CWE-200, CWE-319

src/line/ssrf-guard.ts:83

if (a === 169 && b === 254) return true; // link-local + cloud metadata (169.254.169.254)

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

MEDIUMFilesystem / path · mcp.destructive_tools · CWE-22, CWE-59

line_delete_audience, line_delete_product, line_delete_rich_menu

Why it matters. 3 tool(s) can delete or overwrite

Fix. prefer a read-only mode or scoped tokens; the page states the blast radius

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/imagehost/providers/handoff.ts:19

import { IMAGEMAP_SIZES } from "../../constants.js";

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/imagehost/providers/local-tunnel.ts:40

import { IMAGEHOST_DEFAULT_KEEPALIVE_S, IMAGEMAP_SIZES, TUNNEL_START_TIMEOUT_MS } from "../../constants.js";

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/imagehost/providers/self.ts:16

import { IMAGEHOST_DEFAULT_KEEPALIVE_S, IMAGEMAP_SIZES } from "../../constants.js";

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

src/richmenu/fonts.ts:53

const ASSET_FONT_DIR = fileURLToPath(new URL("../../assets/fonts/", import.meta.url));

LOWFilesystem / path · fs.traversal · CWE-22, CWE-59

tests/guardrails/multi-oa.test.ts:23

} from "../../src/config/multi-oa.js";

LOWNetwork egress · net.metadata_ip · CWE-200, CWE-319

tests/security/ssrf-guard.test.ts:31

return [{ address: "169.254.169.254", family: 4 }];

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

LOWNetwork egress · net.metadata_ip · CWE-200, CWE-319

tests/security/ssrf-guard.test.ts:58

"169.254.169.254", // cloud metadata!

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

LOWNetwork egress · net.metadata_ip · CWE-200, CWE-319

tests/security/ssrf-guard.test.ts:98

expect(isPrivateAddress("::ffff:169.254.169.254")).toBe(true);

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

LOWNetwork egress · net.metadata_ip · CWE-200, CWE-319

tests/security/ssrf-guard.test.ts:129

"https://169.254.169.254/latest/meta-data/",

Why it matters. cloud metadata endpoint: the classic SSRF credential grab

LOWNetwork egress · net.raw_ip · CWE-200, CWE-319

docs/http-transport.md:31

[line-oa-mcp-ultimate] Ready (http) — http://127.0.0.1:3000/mcp

LOWNetwork egress · net.raw_ip · CWE-200, CWE-319

docs/http-transport.md:32

Health: http://127.0.0.1:3000/health

LOWNetwork egress · net.raw_ip · CWE-200, CWE-319

docs/http-transport.md:38

curl http://127.0.0.1:3000/health

LOWNetwork egress · net.raw_ip · CWE-200, CWE-319

docs/http-transport.md:85

curl -s -X POST "http://127.0.0.1:3000/mcp" \

LOWNetwork egress · net.raw_ip · CWE-200, CWE-319

docs/http-transport.md:99

4. Remote MCP server URL: `http://127.0.0.1:3000/mcp`

LOWObfuscation / stealth · obf.base64_blob · CWE-506, CWE-94

tests/imagehost-resize.test.ts:42

"/9j/4AAQSkZJRgABAQAAAQABAAD/2wBDAAIBAQEBAQIBAQECAgICAgQDAgICAgUEBAMEBgUGBgYFBgYGBwkIBgcJBwYGCAsICQoKCgoKBggLDAsKDAkKCgr/2wBDAQICAgICAgUDAwUKBwYHCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgoKCgo

LOWSupply chain · supply.unpinned · CWE-829, CWE-1357

package.json

@modelcontextprotocol/sdk, @resvg/resvg-js, satori, zod, @types/node, @typescript-eslint/eslint-plugin, @typescript-eslint/parser, @vitest/coverage-v8

Why it matters. 14 dependency range(s) float

Fix. pin exact versions or ship a lockfile

LOWPrompt injection · prompt.transfer_instruction · CWE-94, CWE-1427

docs/v2-prd.md:67

**CUT (verified ไม่มี API / ขัด token-only):** auto-reply rules, greeting setter, chat tags, get-content (ต้อง inbound messageId), reply (replyToken), markAsRead (markAsReadToken), standalone "Rich Vi

Why it matters. an instruction to move sensitive data to an outside destination

Fix. remove; a skill never needs the user's secrets off the machine

INFOInventory / provenance · inv.oversize · CWE-1104

docs/brand/logo-v22-tagline.png

docs/brand/logo-v22-tagline.png

Why it matters. 1081980 bytes not read

INFOInventory / provenance · inv.oversize · CWE-1104

docs/brand/logo-v221-tagline.png

docs/brand/logo-v221-tagline.png

Why it matters. 1063396 bytes not read

INFOPrompt injection · prompt.credential_read · CWE-94, CWE-1427

docs/myshop-guide-th.md:34

| กุญแจที่ใช้ | Channel Access Token | **MyShop API Key** (คนละตัว) |

Why it matters. asks the agent to read credentials

Gates applied: no_behavioural_pass.

Audited 2026-10-07 · audit v0.4.1 · source sha 90dcad810469full audit observations/trust-audit/mcp-server/wasintoh__line-oa.json · Report an issue / request a re-scan

Audit history

Every audit this server has had. A grade with a past is a grade somebody is still checking.

Date	Source	Verdict	Grade	Score	Change
2026-10-07	`90dcad810469`	CAUTION	B	86	first audit

Questions

What is the LINE OA MCP server?

What tools does LINE OA expose?

51 in total: 33 read-only, 15 that write, and 3 that can delete or overwrite (line_delete_audience, line_delete_product, line_delete_rich_menu). Every one is listed on this page with its risk.

Is LINE OA safe to connect to an agent?

With care. The audit graded it B (86/100) and found 22 things worth knowing before you trust this server, listed below with the exact line each was found on. Separately from the audit: 3 of its tools can destroy data, so scope the token you give it to what you actually need.

What credentials does LINE OA need?

It reads LINE_CHANNEL_ACCESS_TOKEN, LINE_CHANNEL_SECRET, LINE_MYSHOP_API_KEY and MCP_HTTP_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.

How does LINE OA run?

It speaks stdio and streamable-http, so it runs as a local process your client starts. It is published on npm as line-oa-mcp-ultimate at 2.2.2.

How current is this page?

The grade is for one exact copy of the source (90dcad810469), read on 2026-10-07. The repository is watched and re-audited when it changes.

Source	wasintoh/line-oa-mcp-ultimate
npm	`line-oa-mcp-ultimate@2.2.2`
Transports	stdio · streamable-http
Credentials it reads	`LINE_CHANNEL_ACCESS_TOKENLINE_CHANNEL_SECRETLINE_MYSHOP_API_KEYMCP_HTTP_TOKEN`
License	MIT
Activity	pushed 43d ago