Atlas / MCP servers / nandarona-inc / rootpublish-ai-facts

rootpublish-ai-factsCAUTION

mcp/nandarona-inc/rootpublish-ai-facts

Claude plugin: check what AI assistants tell buyers about your company against your own pricing pages. No API key.

Verdict
CAUTION
Grade
B
Trust score
83 /100
Exposed tools
4 3r · 1w · 0d
Transport
streamable-http
License
MIT
Stars
0
01

Overview

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

rootpublish-ai-facts

Buyers ask AI assistants what a vendor costs before they talk to it. This plugin asks three questions a buyer asks about your company, each of a fresh assistant with web search, and compares the answers with your own pricing and service pages. The report shows every statement that differs from your pages, the facts the assistants stated that your pages do not, the pages a wrong figure came from, and what changed since your last check.

In September 2026 we asked an AI assistant with web search about 30 B2B companies that publish their pricing. For 12 of them it stated a price, plan, contract term or feature that contradicts the company's own pricing page (details).

It runs on your own Claude: no API key, and nothing is sent to Rootpublish.

Install

In Claude Code:

/plugin marketplace add nandarona-inc/rootpublish-ai-facts
/plugin install rootpublish-ai-facts@rootpublish

Then run the check with your site:

/rootpublish-ai-facts:ai-facts-check https://example.com

Requires Node.js, which comes with Claude Code. See the plugin's README for how it works and its limits.

MIT License. Nandarona Inc. / Rootpublish — https://rootpublish.com/ai-facts — [email protected]

Read from source at commit 50fa8508aad8OBSERVED · 2026-10-11
02

Exposed tools (4)

3 read · 1 write · 0 destructive.

ToolRiskDescription
finish_checkreadKeep only the judgements whose quotes are really in the answer and on the company
judge_promptreadFor a fact-judge agent: the prompt that compares one recorded answer with the company
record_answersreadRecord the buyer agents
start_checkwriteStart an AI facts check for a company
03

Trust audit

CAUTIONgrade B · trust 83/100 Install with care. The audit found things worth knowing before you trust its output.

LayerWhat it checksResult
L0Provenance & inventoryPASS
L1Static analysis of the codeWARN
L2Instruction surface (what it tells the agent)PASS
L3Class-specific surfaceWARN
L4Behavioural (sandbox)SKIPPED

What the source does

Filesystem
none-observed
Network
declared (3 observation(s))
Shell
none-observed
Dependencies
pinned
Secrets in source
none-found

Findings (9)

MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:7137
return `ورودی نامعتبر: میبایست instanceof ${issue2.expected} میبود، ${received} دریافت شد`;
MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:7139
return `ورودی نامعتبر: میبایست ${expected} میبود، ${received} دریافت شد`;
MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:7143
return `ورودی نامعتبر: میبایست ${stringifyPrimitive(issue2.values[0])} میبود`;
MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:7145
return `گزینه نامعتبر: میبایست یکی از ${joinValues(issue2.values, "|")} میبود`;
MEDIUMObfuscation / stealth · obf.zero_width · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:9053
file: { unit: "ಬೈಟ್ಗಳು", verb: "ಹೊಂದಲು" },
MEDIUMAuth / authz · mcp.remote_no_auth · CWE-287, CWE-862
streamable-http
Why it matters. a network transport with no auth environment variable found
Fix. require a token
LOWObfuscation / stealth · obf.decode_call · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:1316
const binaryString = atob(base64);
LOWObfuscation / stealth · obf.decode_call · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:2998
atob(data);
LOWObfuscation / stealth · obf.decode_call · CWE-506, CWE-94
rootpublish-ai-facts/server/facts.mjs:3120
const parsedHeader = JSON.parse(atob(header));

Gates applied: no_behavioural_pass.

Audited 2026-10-11 · audit v0.4.1 · source sha 50fa8508aad8full audit observations/trust-audit/mcp-server/nandarona-inc__rootpublish-ai-facts.json · Report an issue / request a re-scan
04

Audit history

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

DateSourceVerdictGradeScoreChange
2026-10-1150fa8508aad8CAUTIONB83first audit
05

Questions

What is the rootpublish-ai-facts MCP server?

Claude plugin: check what AI assistants tell buyers about your company against your own pricing pages. No API key.

What tools does rootpublish-ai-facts expose?

4 in total: 3 read-only, 1 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.

Is rootpublish-ai-facts safe to connect to an agent?

With care. The audit graded it B (83/100) and found 9 things worth knowing before you trust this server, listed below with the exact line each was found on.

What credentials does rootpublish-ai-facts need?

No credential environment variables were found in its source, so it appears to need none.

How does rootpublish-ai-facts run?

It speaks streamable-http, so it runs as a service you connect to over the network.

How current is this page?

The grade is for one exact copy of the source (50fa8508aad8), read on 2026-10-11. The repository is watched and re-audited when it changes.

Advertisement