MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server
dev.workers.b2bintel-edge.sec-biotech-priorart-gateway/mcp-tool-poisoning-prompt-injection-eu-ai-act-mcpsec-biotech-priorart-gateway.b2bintel-edge.workers.devstreamable-httpv2.4.0Scanned
Scan external MCP servers for tool-poisoning, hidden Unicode injection & EU AI Act Annex III.
6 tools: 4 change things and 2 only read.
- Reads private data · 0
- Reads text others write · 6
- Can send data out · 0
- 1 finding
Tools listed without sign-in. Remote posture
With Anlyon
4 tools here change things. Govern them.
- Anlyon lists the tools and adopts none.
- Each tool you adopt gets approvals and receipts.
- Your agent never holds the key.
A tool your agent calls directly is not routed through Anlyon and is not governed by it. Anlyon Radar only lists and labels, and these labels are the same either way.
What the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server can do
The MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server is published by sec-biotech-priorart-gateway.b2bintel-edge.workers.dev, a domain the MCP Registry verified. It has 6 tools. It can create or change data (4 tools: audit_package_cve_and_cisa_kev, batch_sbom_cve_audit_x402, optimize_compute_energy_grid_dispatch and audit_mcp_tool_poisoning_and_eu_ai_act), but cannot delete data. 2 tools only read.
6 tools bring in text other people can write (audit_package_cve_and_cisa_kev, search_patent_academic_prior_art, track_biotech_clinical_trials and 3 more).
The fixed checks matched 1 finding (credentials-mention). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 21, 2026, version 2.4.0.
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.
6 tools
Changes things 4Claude Code: ask
audit_mcp_tool_poisoning_and_eu_ai_actAudits any external MCP tools/list schema for hidden prompt-injection, credential exfiltration, and Unicode steganography, plus 2026 EU AI Act Annex III compliance.Reads text others write1 finding
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.
Findings
- Info
The description mentions "credential".
What this check looks forMatched text
description: "...ist schema for hidden prompt-injection, credential exfiltration, and Unicode steganography..."
Description, as the server wrote it Audits any external MCP tools/list schema for hidden prompt-injection, credential exfiltration, and Unicode steganography, plus 2026 EU AI Act Annex III compliance.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Scan Remote MCP Server for Tool-Poisoning / Prompt-Injection & Classify EU AI Act Risk
audit_package_cve_and_cisa_kevAudits an npm, PyPI, Go, Cargo, or Maven package version against OSV.dev and US CISA Known Exploited Vulnerabilities.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 Audits an npm, PyPI, Go, Cargo, or Maven package version against OSV.dev and US CISA Known Exploited Vulnerabilities.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Audit Software Package for OSV CVEs & CISA Exploited Vulnerabilities
batch_sbom_cve_audit_x402Audits up to 15 dependency specs in parallel against OSV.dev + CISA KEV ($0.006 USDC on Base via x402).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 Audits up to 15 dependency specs in parallel against OSV.dev + CISA KEV ($0.006 USDC on Base via x402).
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Batch Audit Up to 15 SBOM Dependencies ($0.006 USDC via x402)
optimize_compute_energy_grid_dispatchRanks 6 global datacenter and robotics charging grids (ERCOT, PJM, CAISO, Frankfurt, UK, Seoul) by live solar/wind surplus, cooling PUE, and power cost.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 Ranks 6 global datacenter and robotics charging grids (ERCOT, PJM, CAISO, Frankfurt, UK, Seoul) by live solar/wind surplus, cooling PUE, and power cost.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Optimize AI Compute & Robotics Fleet Dispatch Across Global Power Grids
Only reads 2Show 2 read-only tools
search_patent_academic_prior_artSearches 250M+ technical works for patent novelty and prior-art assessment, returning reconstructed abstracts, DOIs, and citation counts.Reads text others writeNo findings
Why it is in this group
- The name "search_patent_academic_prior_art" says "search": a word for reading.
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 Searches 250M+ technical works for patent novelty and prior-art assessment, returning reconstructed abstracts, DOIs, and citation counts.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Search Global Patent & Scholarly Prior-Art Works
track_biotech_clinical_trialsQueries ClinicalTrials.gov API v2 for active Phase 1-3 trials, sponsors, NCT IDs, and completion dates.Reads text others writeNo findings
Why it is in this group
- The description's first verb is "query": a word for reading.
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 Queries ClinicalTrials.gov API v2 for active Phase 1-3 trials, sponsors, NCT IDs, and completion dates.
Annotations the server declares
- readOnlyHint: true
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- title: Track Live ClinicalTrials.gov Biotech & MedTech Pipelines
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 or outbound.
- Reads private data0
- Reads files, mail, messages, databases, secrets, calendars or repositories.
- Reads text others write6
- Brings in text someone else can write: web pages, inbound email, messages, issues and comments.
- Can send data out0
- 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 21, 2026, 19:25 UTC, unchanged since Nov 21, 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
- 3
- Protocol version
- 2025-03-26
- WWW-Authenticate on 401
- not observed
- resource_metadata in WWW-Authenticate
- not observed
- RFC 9728 protected resource metadata
- not observed
- Authorization servers
- not observed
- Legacy HTTP+SSE transport
- no
- Dynamic client registration
- not observed
- Client ID metadata documents
- not observed
- WarnThe whole server
The endpoint served tools/list without credentials, and the list has 3 tools that change things (3 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://sec-biotech-priorart-gateway.b2bintel-edge.workers.dev/mcp: tools/list returned 6 tools to an anonymous client, including audit_package_cve_and_cisa_kev (write), batch_sbom_cve_audit_x402 (write), audit_mcp_tool_poisoning_and_eu_ai_act (write)
Findings
Matches of fixed checks, with the text they matched.
The description mentions "credential".
What this check looks forMatched text
description: "...ist schema for hidden prompt-injection, credential exfiltration, and Unicode steganography..."
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 2.4.0. Values in angle brackets are yours to fill in.
claude mcp add --transport http tool-poisoning-prompt-injection-eu-ai-act https://sec-biotech-priorart-gateway.b2bintel-edge.workers.dev/mcpclaude mcp add --transport http tool-poisoning-prompt-injection-eu-ai-act https://sec-biotech-priorart-gateway.b2bintel-edge.workers.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": [
"mcp__tool-poisoning-prompt-injection-eu-ai-act__search_patent_academic_prior_art",
"mcp__tool-poisoning-prompt-injection-eu-ai-act__track_biotech_clinical_trials"
],
"ask": [
"mcp__tool-poisoning-prompt-injection-eu-ai-act__audit_package_cve_and_cisa_kev",
"mcp__tool-poisoning-prompt-injection-eu-ai-act__batch_sbom_cve_audit_x402",
"mcp__tool-poisoning-prompt-injection-eu-ai-act__optimize_compute_energy_grid_dispatch",
"mcp__tool-poisoning-prompt-injection-eu-ai-act__audit_mcp_tool_poisoning_and_eu_ai_act"
],
"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 tool-poisoning-prompt-injection-eu-ai-act. Review it before you use it.
Capability tags behind 6 rules
- mcp__tool-poisoning-prompt-injection-eu-ai-act__audit_package_cve_and_cisa_kev untrusted-input
- mcp__tool-poisoning-prompt-injection-eu-ai-act__search_patent_academic_prior_art untrusted-input
- mcp__tool-poisoning-prompt-injection-eu-ai-act__track_biotech_clinical_trials untrusted-input
- mcp__tool-poisoning-prompt-injection-eu-ai-act__batch_sbom_cve_audit_x402 untrusted-input
- mcp__tool-poisoning-prompt-injection-eu-ai-act__optimize_compute_energy_grid_dispatch untrusted-input
- mcp__tool-poisoning-prompt-injection-eu-ai-act__audit_mcp_tool_poisoning_and_eu_ai_act untrusted-input
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/dev.workers.b2bintel-edge.sec-biotech-priorart-gateway/mcp-tool-poisoning-prompt-injection-eu-ai-act-mcp/verify?hash=c1844f279923088029900a8569f8c591ba3ed6be0b36d413dcc4d1c505be5b74" | jq -e '.changed == false'Pinned hash c1844f279923, from Nov 21, 2026, 07:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/dev.workers.b2bintel-edge.sec-biotech-priorart-gateway/mcp-tool-poisoning-prompt-injection-eu-ai-act-mcp/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/dev.workers.b2bintel-edge.sec-biotech-priorart-gateway/mcp-tool-poisoning-prompt-injection-eu-ai-act-mcp)Source: MCP Registry · repository. Synced Nov 21, 2026, 07:17 UTC.
MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server safe?
- No scan can promise that. What its definitions show: 6 tools, 4 of which can change, send or delete something, and 1 finding from fixed checks. Allow the reads, ask before each change, and deny the deletes; the permissions block on this page does that.
Can the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server do things that cannot be undone?
- No tool is classed as irreversible: none of them sends, pays, publishes or merges, going by their definitions.
What tools does the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server have?
- 6 tools: 4 write, 2 read. audit_package_cve_and_cisa_kev, batch_sbom_cve_audit_x402, optimize_compute_energy_grid_dispatch, audit_mcp_tool_poisoning_and_eu_ai_act, search_patent_academic_prior_art and track_biotech_clinical_trials.
Does the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server send data to outside services?
- None of its tools sends, posts or publishes, or takes a recipient, channel or URL.
Has the MCP Tool-Poisoning, Prompt-Injection & 2026 EU AI Act Auditor MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 21, 2026.