AImart Parquet MCP server
io.github.balogvn/aimart-parquetbalogvnstreamable-httpv0.1.3Scanned
Messy CSV, JSON, Excel, SPSS/Stata/SAS and PDF tables to typed Parquet; inspect, query, join.
6 tools: 4 change things and 2 only read.
- Reads private data · 3
- Reads text others write · 4
- Can send data out · 4
- 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 AImart Parquet MCP server can do
The AImart Parquet MCP server is published by the GitHub account balogvn. It has 6 tools. It can create or change data (4 tools: convert_to_parquet, query_parquet, verify_provenance and request_capability), but cannot delete data. 2 tools only read.
3 tools read private data (inspect_parquet, preview_parquet and query_parquet); 4 tools bring in text other people can write (convert_to_parquet, inspect_parquet, preview_parquet and 1 more); 4 tools can send data out of your system (convert_to_parquet, inspect_parquet, preview_parquet and 1 more). On its own it has all three parts of a prompt-injection exfiltration path: private data, text others write, and a way out.
The fixed checks matched 1 finding (annotation-conflict). Each one comes with the text it matched.
Its remote endpoint lists its tools without sign-in.
Last scanned Nov 17, 2026, version 0.1.3.
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
convert_to_parquetConvert a messy data file into clean, typed, zstd-compressed Parquet.Reads text others writeCan send data outNo findings
Why it is in this group
- The name "convert_to_parquet" says "convert": a word for a change.
What it can reach
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Description, as the server wrote it Convert a messy data file into clean, typed, zstd-compressed Parquet. Use this first for any CSV/TSV, JSON/NDJSON, Excel, SPSS, Stata or SAS file, or a digital PDF with tables. It detects the format, encoding and delimiter; skips title rows; repairs broken JSON; pulls tables out of PDFs (merging tables that continue across pages); infers int/float/bool/date/timestamp columns (including 1.234,56 and $1,234 formats) while keeping IDs with leading zeros as text; normalizes column names to snake_case; keeps SPSS/Stata/SAS variable and value labels; and can validate against an expected_schema. Returns an artifact_id (pass it to inspect_parquet / preview_parquet / query_parquet), a download link, and a per-column report: type, null count, values that did not fit the chosen type, and warnings. Price: $0.02 USDC per successful call, paid via x402.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: true
- readOnlyHint: false
query_parquetRun one read-only DuckDB SELECT over a Parquet file, available as table `data`; join more files by passing `tables` ({name: artifact_id}, up to 7 besides data), e.g.Reads private dataReads text others writeCan send data out1 finding
Why it is in this group
- The name "query_parquet" says "query": a word for reading.
- The description's first verb is "run": a word for a change.
- The input takes a statement or command to run (input.sql), which can change anything it can reach.
- The server marks it read-only (readOnlyHint: true), but an annotation never lowers a class.
What it can reach
- private-dataThe description's first sentence says "db": it reads databases.
- private-dataThe input names a table (input.tables), and the tool reads: it reads databases.
- private-dataThe name says "query" and the input takes a statement (input.sql): it reads databases.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Findings
- Warn
The server marks "query_parquet" 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 "query_parquet" says "query": a word for reading.
Description, as the server wrote it Run one read-only DuckDB SELECT over a Parquet file, available as table `data`; join more files by passing `tables` ({name: artifact_id}, up to 7 besides data), e.g. a GL against a bank statement. Supports aggregates, window functions, CTEs, DESCRIBE data and SUMMARIZE data. Up to 1000 rows are returned; `truncated` says if more matched. Writes, file access and multiple statements are rejected. Example: SELECT region, sum(units) FROM data GROUP BY 1 ORDER BY 2 DESC. Price: $0.005 USDC per successful call, paid via x402.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
request_capabilityFree.No findings
Why it is in this group
- The name "request_capability" says "request": a word for a change.
Description, as the server wrote it Free. If none of these tools does what you need — a file type we reject, an output we do not produce, a missing operation — describe the task, the formats involved, and the most you would pay per call. The operator reads these to decide what to build next. Do not include personal or confidential data.
Annotations the server declares
- destructiveHint: false
- idempotentHint: false
- openWorldHint: false
- readOnlyHint: false
verify_provenanceFree.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 Free. Check that a file this service produced is authentic and unaltered: verifies the post-quantum ML-DSA-65 (NIST FIPS 204) signature in an artifact's `provenance`, and with artifact_id also re-checks the stored file's hash and size against the signed manifest.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: false
- readOnlyHint: true
Only reads 2Show 2 read-only tools
inspect_parquetDescribe a Parquet file from its footer without reading rows: row count, columns, types, null counts, min/max per column, and SPSS/Stata/SAS labels.Reads private dataReads text others writeCan send data outNo findings
Why it is in this group
- The name "inspect_parquet" says "inspect": a word for reading.
What it can reach
- private-dataThe description's first sentence says "describe" and "file": it reads files.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Claude Code: ask, because it is tagged outbound.
Description, as the server wrote it Describe a Parquet file from its footer without reading rows: row count, columns, types, null counts, min/max per column, and SPSS/Stata/SAS labels. The cheapest way to understand a dataset before querying it. Input: artifact_id from convert_to_parquet, or a url to a .parquet file. Price: $0.002 USDC per successful call, paid via x402.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: true
preview_parquetReturn a slice of rows from a Parquet file (default 20, max 500), optionally only some columns, as JSON rows, a Markdown table, or CSV.Reads private dataReads text others writeCan send data outNo findings
Why it is in this group
- The name "preview_parquet" says "preview": a word for reading.
What it can reach
- private-dataThe description's first sentence says "return" and "file": it reads files.
- untrusted-inputThe server marks it openWorldHint: true: it reaches things outside the server's own data.
- untrusted-inputThe input takes a URL (input.url), and the tool reads it: it reads web pages, which anyone can write.
- outboundThe input takes a URL (input.url): a request to a URL can carry data to whoever runs it.
Claude Code: ask, because it is tagged outbound.
Description, as the server wrote it Return a slice of rows from a Parquet file (default 20, max 500), optionally only some columns, as JSON rows, a Markdown table, or CSV. Use to eyeball data; use query_parquet to filter or aggregate. Long cells are truncated. Price: $0.002 USDC per successful call, paid via x402.
Annotations the server declares
- destructiveHint: false
- idempotentHint: true
- openWorldHint: true
- readOnlyHint: 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.
48 paths in this server alone from private data, through text someone else wrote, to a way out. Denying 3 private-data tools breaks every one.
- Reads private data3
- 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 out4
- Can carry data to people or systems: it sends, posts or publishes, or takes a recipient, channel or URL.
{
"permissions": {
"deny": [
"mcp__aimart-parquet__inspect_parquet",
"mcp__aimart-parquet__preview_parquet",
"mcp__aimart-parquet__query_parquet"
]
}
}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 19, 2026, 17:21 UTC, unchanged since Nov 12, 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
- 4
- Protocol version
- 2025-06-18
- 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 4 tools that change things (4 write). Whether a call also works without credentials was not tested: tools are never called.
What this check looks forMatched text
https://balogvn--aimart-parquet.us-east.modal.run/mcp: tools/list returned 6 tools to an anonymous client, including convert_to_parquet (write), query_parquet (write), verify_provenance (write), request_capability (write)
Findings
Matches of fixed checks, with the text they matched.
- Warnquery_parquet
The server marks "query_parquet" 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 "query_parquet" says "query": a word for reading.
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 0.1.3. Values in angle brackets are yours to fill in.
claude mcp add --transport http aimart-parquet https://balogvn--aimart-parquet.us-east.modal.run/mcpclaude mcp add --transport http aimart-parquet https://balogvn--aimart-parquet.us-east.modal.run/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__aimart-parquet__convert_to_parquet",
"mcp__aimart-parquet__inspect_parquet",
"mcp__aimart-parquet__preview_parquet",
"mcp__aimart-parquet__query_parquet",
"mcp__aimart-parquet__verify_provenance",
"mcp__aimart-parquet__request_capability"
],
"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 aimart-parquet. Review it before you use it.
Capability tags behind 4 rules
- mcp__aimart-parquet__convert_to_parquet untrusted-input, outbound
- mcp__aimart-parquet__inspect_parquet private-data, untrusted-input, outbound
- mcp__aimart-parquet__preview_parquet private-data, untrusted-input, outbound
- mcp__aimart-parquet__query_parquet private-data, 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.balogvn/aimart-parquet/verify?hash=3c2cdfd2418aac4156025049f7c4b336da5fefeeffe1fe44498ceb095171ba1e" | jq -e '.changed == false'Pinned hash 3c2cdfd2418a, from Nov 17, 2026, 19:18 UTC.
Watch
Get told when its tools change: a new tool, a new description, a new class.
https://radar.anlyon.com/servers/io.github.balogvn/aimart-parquet/changes.xmlAn Atom feed. Works in any feed reader or Slack's RSS app.
Badge
[](https://radar.anlyon.com/servers/io.github.balogvn/aimart-parquet)Source: MCP Registry · repository. Synced Nov 17, 2026, 19:17 UTC.
AImart Parquet MCP server: questions and answers
Answered from the scan above, not from the publisher's description.
Is the AImart Parquet 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 AImart Parquet MCP server delete data?
- No tool is classed as destructive: going by their definitions, none of them deletes, cancels or revokes anything.
Can the AImart Parquet 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 AImart Parquet MCP server have?
- 6 tools: 4 write, 2 read. convert_to_parquet, query_parquet, verify_provenance, request_capability, inspect_parquet and preview_parquet.
Does the AImart Parquet MCP server need sign-in or OAuth?
- Its remote endpoint listed its tools to an anonymous client, without sign-in.
Does the AImart Parquet MCP server send data to outside services?
- 4 tools can carry data out (it sends, posts or publishes, or takes a recipient, channel or URL): convert_to_parquet, inspect_parquet, preview_parquet and query_parquet.
Has the AImart Parquet MCP server changed its tools?
- Not since Anlyon Radar started tracking it. Last scanned Nov 17, 2026.