HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server
io.github.ogasurfproject-jpg/horizon-shield-webmcpogasurfproject-jpgstreamable-httpv1.0.4Scanned
Agent intake desk for property renovation. Read only, CORS open, fail closed, receipts recompute.
5 tools: 1 do something that can't be undone and 4 change things.
- Reads private data · 0
- Reads text others write · 4
- Can send data out · 1
- 3 findings
Tools listed without sign-in. Remote posture
Anlyon alternative
5 of these tools change something. Run those as governed actions.
Declare what an action changes, and Anlyon puts it under a shared limit, binds every approval to the exact request, and grades the outcome. It works on any HTTPS API you can describe as an action.
Anlyon Radar does not run, proxy or govern this MCP server. A tool your own code calls directly is not routed through Anlyon and is not governed by it.
Import an OpenAPI spec →What the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server can do
The HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server is published by the GitHub account ogasurfproject-jpg. It has 5 tools. It can do things that cannot be undone, such as sending, paying or publishing (draft_broadcast) and create or change data (4 tools: route_request, run_full_audit, intake_estimate and scan_tactics), but cannot delete data. None of its tools only reads.
4 tools bring in text other people can write (route_request, run_full_audit, scan_tactics and 1 more); 1 tool can send data out of your system (draft_broadcast).
The fixed checks matched 3 findings (annotation-conflict). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 21, 2026, version 1.0.4.
Tools
Grouped by what each tool does, most consequential first. Open one for its reasons and full description; descriptions are the publisher's own words.
5 tools
Can't be undone 1Claude Code: ask
draft_broadcastある工事の過剰請求への注意喚起を発信する下書き(note用長文・X用短文)を生成し、HORIZON SHIELDの該当解説ページ(実在URL)への被リンクを添える。価格はKIRA(検証可能SHA-256付き)の一次データのみ。推測の数字は入れない。下書きであり公開前に運営者が最終版にする(自動投稿しない)。 / Generates broadcast DRA…Reads text others writeCan send data out1 finding
Why it is in this group
- The name "draft_broadcast" says "broadcast": a word for sending, and what is sent cannot be taken back.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe name "draft_broadcast" says "broadcast": it sends or publishes something to other people.
Findings
- Warn
The server marks "draft_broadcast" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "draft_broadcast" says "broadcast": a word for sending, and what is sent cannot be taken back.
Description, as the server wrote it ある工事の過剰請求への注意喚起を発信する下書き(note用長文・X用短文)を生成し、HORIZON SHIELDの該当解説ページ(実在URL)への被リンクを添える。価格はKIRA(検証可能SHA-256付き)の一次データのみ。推測の数字は入れない。下書きであり公開前に運営者が最終版にする(自動投稿しない)。 / Generates broadcast DRAFTS with backlinks. Verifiable first-party prices only. Draft; operator finalizes. No auto-posting. USE WHEN: the operator wants note/X draft text warning about overcharging for a given work type. DO NOT USE WHEN: the caller wants a verdict on a specific quote (call intake_estimate) or the tactic list itself (call scan_tactics). Output is a DRAFT for a human to finalise. This tool never posts.
Annotations the server declares
- title: 注意喚起の発信下書き
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
Changes things 4Claude Code: ask
intake_estimate施主の建設・リフォーム見積もりを受け付け、HORIZON SHIELD KIRA(内部・中立)の適正価格診断へ橋渡しする集客窓口。工事名と業者提示額を渡すと、適正かどうかの判定と、無料の第三者チェック(EHN)への導線を返す。価格の断定はせず、確認すべき点を渡す。 / Intake desk: bridges a homeowner quote to the…No findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
Description, as the server wrote it 施主の建設・リフォーム見積もりを受け付け、HORIZON SHIELD KIRA(内部・中立)の適正価格診断へ橋渡しする集客窓口。工事名と業者提示額を渡すと、適正かどうかの判定と、無料の第三者チェック(EHN)への導線を返す。価格の断定はせず、確認すべき点を渡す。 / Intake desk: bridges a homeowner quote to the KIRA fair-price audit and returns the verdict plus a free third-party check (EHN) path. USE WHEN: a homeowner has a specific quoted price for a named construction or renovation work item in Japan and wants to know whether that price is reasonable. DO NOT USE WHEN: there is no price yet (call ask), the caller wants the known overcharge tactics rather than a verdict on their own quote (call scan_tactics), or the caller wants publishable text (call draft_broadcast). Returns a verdict and a free third-party check path. It does not draft or post.
Annotations the server declares
- title: 見積もり適正診断(KIRA)
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
route_requestCheck whether a construction or renovation quote in JAPAN is fairly priced, against the open JCCDB (526,128 records: line items and observations) and Bitcoin-anchored verification …Reads text others write1 finding
Why it is in this group
- The name "route_request" says "request": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Warn
The server marks "route_request" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "route_request" says "request": a word for a change.
Description, as the server wrote it Check whether a construction or renovation quote in JAPAN is fairly priced, against the open JCCDB (526,128 records: line items and observations) and Bitcoin-anchored verification records. Single entry point: pass one sentence. USE WHEN: the user has a quote or a price for building/renovation work in Japan and wants to know if it is reasonable; or wants to check a HORIZON SHIELD receipt or a jidec: citation; or asks about overcharging tactics used by contractors; or is looking for a verified contractor. DO NOT USE WHEN: the work is outside Japan; the question is not about construction or renovation; the user has already signed or paid and needs consumer-protection help; or there is an active emergency (gas smell, collapse, water not stopping). In those cases this tool does not answer - it returns the appropriate outside destination instead, and says so. Every reply carries verify (a URL that checks the answer) and limits (what the answer does NOT prove). Routing is a deterministic keyword table, not an LLM, so the same input always routes the same way. Examples: "is 800,000 yen high for exterior wall painting?" / "is jidec:entry:9 genuine?" / "how do I refuse a door-to-door sales pitch?" 日本の建設・リフォーム見積もりが適正かを確認する単一入口。日本語の一文で渡してよい。適正診断・台帳(JIDEC)の記録検証・過剰請求の手口・検証済み加盟店の4方向へ決定的に振り分ける。日本国外/建設以外/契約後の紛争/緊急時は**答えずに適切な外部の窓口を返す。**返り値には必ず verify(検証URL)と limits(証明していないこと)が入る。
Annotations the server declares
- title: 一文で聞けば、正しい窓口に繋ぐ
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
run_full_audit1回の呼び出しで HORIZON SHIELD の集客→診断→注意喚起→発信を一気通貫で回す司令塔。work(と任意の quoted_price)を渡すと、内部で intake(KIRA適正診断)・scan_tactics(検証済み手口+一次ソース)・draft_broadcast(発信下書き+被リンク)を順に実行し、結果を1つに束ねて返す。価格は検証可能な…Reads text others write1 finding
Why it is in this group
- The name "run_full_audit" says "run": a word for a change.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Findings
- Warn
The server marks "run_full_audit" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_full_audit" says "run": a word for a change.
Description, as the server wrote it 1回の呼び出しで HORIZON SHIELD の集客→診断→注意喚起→発信を一気通貫で回す司令塔。work(と任意の quoted_price)を渡すと、内部で intake(KIRA適正診断)・scan_tactics(検証済み手口+一次ソース)・draft_broadcast(発信下書き+被リンク)を順に実行し、結果を1つに束ねて返す。価格は検証可能な一次データのみ。発信は下書きで自動投稿しない。 / One-call orchestrator returning audit + tactics + broadcast draft. Verifiable first-party prices only. Drafts only, no auto-posting. USE WHEN: the caller wants all three steps (audit, tactics, broadcast draft) for one named work item in a single call and will act on the combined result. DO NOT USE WHEN: only one step is needed. Call intake_estimate for a verdict on a specific quote, scan_tactics for the documented overcharge tactics, or draft_broadcast for publishable text. If the request is still a free-form question with no work item, call ask first. This tool never posts anything.
Annotations the server declares
- title: 集客→診断→発信 一括実行
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
scan_tacticsある工事・キーワードに関する『過剰請求の手口』を、HORIZON SHIELD(大賀俊勝30年監修)の検証済みデータ(内部KIRA)から返し、一次ソース(国民生活センター/消費者庁/EHN実例ボード)の在処を添える。価格判定ではなく注意喚起。推測で新事例を断定しない。 / Returns verified overcharge tactics and poi…Reads text others writeNo findings
Why it is in this group
- Nothing in its name or description says what it does, so it is counted as a write.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
Description, as the server wrote it ある工事・キーワードに関する『過剰請求の手口』を、HORIZON SHIELD(大賀俊勝30年監修)の検証済みデータ(内部KIRA)から返し、一次ソース(国民生活センター/消費者庁/EHN実例ボード)の在処を添える。価格判定ではなく注意喚起。推測で新事例を断定しない。 / Returns verified overcharge tactics and points to primary sources. Awareness, not a price verdict. USE WHEN: the caller wants the documented overcharge tactics for a work type or keyword, together with their primary sources. DO NOT USE WHEN: the caller has their own quote and wants a verdict on it (call intake_estimate), or wants publishable text (call draft_broadcast). This returns awareness material and never a price verdict on a specific quote.
Annotations the server declares
- title: 過剰請求の手口スキャン
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
This is not a safety rating, and no finding is not a safety guarantee. The groups and findings come from fixed checks of the tool definitions the server returned on the date shown, not from its code or what it does when called. How the checks work.
Exfil surface
Which tools read private data, which read text someone else can write, and which can send data out. One of each is a path a prompt injection can use.
No private-data, untrusted-input, outbound path in this server alone: no tool is tagged private-data.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write4
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out1
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
Other servers in the same agent add their own tools to these paths. Audit your whole setup to see them together.
Remote posture
What anyone on the internet sees at the remote endpoint: an anonymous initialize and tools/list, and the public OAuth metadata.
Looked at Nov 20, 2026, 19:23 UTC, unchanged since Oct 22, 2026. No credentials were sent and no tool was called. What is looked at.
- Reachable
- yes
- HTTP status of an anonymous initialize
- 200
- tools/list served without credentials
- yes
- Tools that change things, listed without credentials
- 5
- Protocol version
- 2025-06-18
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- https://web.horizonshield.dev/.well-known/oauth-protected-resource/mcp
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- HighThe whole server
The endpoint served tools/list without credentials, and the list has 5 tools that change things (1 irreversible, 4 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://web.horizonshield.dev/mcp: tools/list returned 5 tools to an anonymous client, including draft_broadcast (irreversible), route_request (write), run_full_audit (write), intake_estimate (write), scan_tactics (write)
Findings
Matches of fixed checks, with the text they matched.
- Warnroute_request
The server marks "route_request" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "route_request" says "request": a word for a change.
- Warnrun_full_audit
The server marks "run_full_audit" read-only (readOnlyHint: true), and its own name, description or input give the class "write".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "run_full_audit" says "run": a word for a change.
- Warndraft_broadcast
The server marks "draft_broadcast" read-only (readOnlyHint: true), and its own name, description or input give the class "irreversible".
What this check looks forMatched text
annotations: {"readOnlyHint":true,"destructiveHint":false,"idempotentHint":true,"openWorldHint":true}; The name "draft_broadcast" says "broadcast": a word for sending, and what is sent cannot be taken back.
History
Changes to its tool definitions and to its remote endpoint, newest first.
No change seen yet. Its tools are rescanned every day, and a remote endpoint is looked at once a week.
Install
Pinned to version 1.0.4. Values in angle brackets are yours to fill in.
claude mcp add --transport http horizon-shield-webmcp https://web.horizonshield.dev/mcpclaude mcp add --transport http horizon-shield-webmcp https://web.horizonshield.dev/mcpRun in a terminal.
Claude Code permissions
Allow the reads, ask before a change or a way out, deny the deletes. Paste into .claude/settings.json.
{
"permissions": {
"allow": [],
"ask": [
"mcp__horizon-shield-webmcp__route_request",
"mcp__horizon-shield-webmcp__run_full_audit",
"mcp__horizon-shield-webmcp__intake_estimate",
"mcp__horizon-shield-webmcp__scan_tactics",
"mcp__horizon-shield-webmcp__draft_broadcast"
],
"deny": []
}
}Tools that only read are allowed, tools that change things ask first, tools that delete or cancel are denied. A read that can send data out (it takes a URL, say) asks first too. It assumes the server is installed as horizon-shield-webmcp. Review it before you use it.
Capability tags behind 4 rules
- mcp__horizon-shield-webmcp__route_request untrusted-input
- mcp__horizon-shield-webmcp__run_full_audit untrusted-input
- mcp__horizon-shield-webmcp__scan_tactics untrusted-input
- mcp__horizon-shield-webmcp__draft_broadcast untrusted-input, outbound
Pin
Keep the decision: fail CI when the tools change, follow every change, or show a badge.
Pin in CI
Fail a build when the tools you reviewed are not the tools the server serves today.
curl -s "https://radar.anlyon.com/api/v1/servers/io.github.ogasurfproject-jpg/horizon-shield-webmcp/verify?hash=85ef5cd46ee032ca9558aaddb677eb6519e7d0930c4bace200a426569d24061c" | jq -e '.changed == false'Pinned hash 85ef5cd46ee0, from Nov 21, 2026, 08:17 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.ogasurfproject-jpg/horizon-shield-webmcp/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.ogasurfproject-jpg/horizon-shield-webmcp)Source: MCP Registry · repository. Synced Oct 21, 2026, 18:42 UTC.
HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server safe?
- No scan can promise that. What its definitions show: 5 tools, 5 of which can change, send or delete something, and 3 findings from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server do things that cannot be undone?
- Yes. 1 tool can send, pay, publish or merge: draft_broadcast.
What tools does the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server have?
- 5 tools: 1 irreversible, 4 write. draft_broadcast, route_request, run_full_audit, intake_estimate and scan_tactics.
Does the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server send data to outside services?
- 1 tool can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): draft_broadcast.
Has the HORIZON SHIELD: Construction Estimate Auditor for Japan (KIRA) MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 21, 2026.