io.github.erichare/astra-mcp

Est-ce que io.github.erichare/astra-mcp est sûr ?

Au dernier audit MCP Hub, le serveur MCP « io.github.erichare/astra-mcp » est classé « Vérifié ». Aucun comportement malveillant détecté au test comportemental. Le verdict est réévalué en continu ; consultez la fiche pour l'état à jour.

Méthode : MCP Hub analyse la description de chaque outil (couche 1) et, pour le verdict fort, exécute le serveur dans une micro-VM isolée avec leurres et réseau bloqué (couche 2). On publie des noms et des compteurs, jamais le contenu brut renvoyé par un outil.

Versions auditées & empreintes

Vérifiez la version que vous utilisez : son empreinte doit correspondre à celle listée ici — sinon vous n'utilisez pas la version auditée.

⚠ La dernière version publiée (2.1.0) n'a pas encore été auditée. Un nouveau pentest est programmé ; les verdicts ci-dessous portent sur les versions listées.

Vous utilisez ce serveur ? Suivez-le (créez un compte) pour être alerté dès qu'une nouvelle version paraît ou que le verdict change.

Les 18 outils

• connection_status≈ 142 tokens

  Check whether Astra DB credentials are configured and working: where the token, endpoint, and keyspace come from (env, .env, plugin settings, Astra CLI), which database is targeted, and a live connectivity check. Call it first when another tool reports not_configured or an auth error.

  Paramètres, retour & exemple (1)

  Paramètres

  • check · boolean · optionnel
    Make a live request to verify the connection (default true).

  Exemple d'appel

  {"name": "connection_status", "arguments": {"check": true}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• list_databases≈ 121 tokens

  List the Astra databases this token can access (DevOps API): name, id, status, regions with Data API endpoints, and keyspaces. Use it to choose the `database` argument of other tools.

  Paramètres, retour & exemple (1)

  Paramètres

  • include · string · optionnel · active | all
    active (default) or all non-terminated databases.

  Exemple d'appel

  {"name": "list_databases", "arguments": {"include": "active"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• database_overview≈ 303 tokens

  Map a database: keyspaces → collections (vector dimension/metric, vectorize model, lexical/rerank, estimated counts) and tables (column and vector-column counts). Start here to see what exists. Renders an interactive overview.

  Paramètres, retour & exemple (5)

  Paramètres

  • emit · string · optionnel · structured | html_file
    html_file also writes a self-contained HTML page and returns a link to it — only for terminal-only hosts, when the user asks for a visual.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspaces · array · optionnel
    Only these keyspaces (default: all).
  • includeCounts · boolean · optionnel
    Estimated document counts per collection (one request each).
  • maxPerKeyspace · integer · optionnel
    Max collections/tables listed per keyspace.

  Exemple d'appel

  {"name": "database_overview", "arguments": {"emit": "structured", "database": "texte", "keyspaces": [], "includeCounts": true, "maxPerKeyspace": 1}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• describe_collection≈ 289 tokens

  Show one collection's configuration and shape: vector settings, vectorize and rerank services, lexical, indexing allow/deny, default id type, estimated count, and a sample document with field types. Use before writing queries or code against it.

  Paramètres, retour & exemple (5)

  Paramètres

  • emit · string · optionnel · structured | html_file
    html_file also writes a self-contained HTML page and returns a link to it — only for terminal-only hosts, when the user asks for a visual.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • collection · string · requis
    Collection name.
  • includeSample · boolean · optionnel
    Include one sample document (vectors omitted).

  Exemple d'appel

  {"name": "describe_collection", "arguments": {"collection": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• describe_table≈ 268 tokens

  Show one table's schema: columns and types, primary key (partition and clustering columns), indexes, vector columns, and sample rows. Use before writing queries or code against it.

  Paramètres, retour & exemple (5)

  Paramètres

  • emit · string · optionnel · structured | html_file
    html_file also writes a self-contained HTML page and returns a link to it — only for terminal-only hosts, when the user asks for a visual.
  • table · string · requis
    Table name.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • includeSample · boolean · optionnel
    Include up to 3 sample rows (vectors omitted).

  Exemple d'appel

  {"name": "describe_table", "arguments": {"table": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• find≈ 550 tokens

  Read documents (collections) or rows (tables) with an optional Data API filter, sort, and projection. Returns pages of 20 plus nextPageState for more. For similarity search use vector_search. Renders an interactive explorer.

  Paramètres, retour & exemple (10)

  Paramètres

  • emit · string · optionnel · structured | html_file
    html_file also writes a self-contained HTML page and returns a link to it — only for terminal-only hosts, when the user asks for a visual.
  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • name · string · requis
    Collection or table name.
  • sort · object · optionnel
    Non-vector sort, e.g. {"createdAt": -1}. For similarity use vector_search.
  • limit · integer · optionnel
    Documents to return (rounded up to whole pages of 20).
  • filter · object · optionnel
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • pageState · string · optionnel
    nextPageState from a previous call, to continue paging.
  • projection · object · optionnel
    Fields to include (1) or exclude (0), e.g. {"title": 1, "author": 1}.

  Exemple d'appel

  {"name": "find", "arguments": {"name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• vector_search≈ 712 tokens

  Similarity search by natural-language `query` (server-side vectorize), an explicit `vector`, or `documentId` (more like this). `hybrid` combines vector and lexical (BM25) results with reranking on collections that have both. A filter narrows candidates. Returns ranked hits with similarity scores; renders an interactive results view.

  Paramètres, retour & exemple (14)

  Paramètres

  • emit · string · optionnel · structured | html_file
    html_file also writes a self-contained HTML page and returns a link to it — only for terminal-only hosts, when the user asks for a visual.
  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • name · string · requis
    Collection or table name.
  • limit · integer · optionnel
    How many hits.
  • query · string · optionnel
    Natural-language text; embedded server-side (vectorize).
  • filter · object · optionnel
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • hybrid · boolean · optionnel
    Collections with lexical + rerank: combined vector + BM25 search, reranked.
  • idType · string · optionnel · auto | string | uuid | objectId
    How to interpret documentId (auto follows the collection's defaultId).
  • vector · array · optionnel
    A query embedding (must match the dimension).
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • documentId · ? · optionnel
    Find items similar to this existing document (collections).
  • projection · object · optionnel
    Fields to include (1) or exclude (0), e.g. {"title": 1, "author": 1}.
  • vectorColumn · string · optionnel
    Tables only: which vector column to search (default: the only one).

  Exemple d'appel

  {"name": "vector_search", "arguments": {"name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• count≈ 267 tokens

  Exact count of documents in a collection matching a filter, up to upperBound (Data API maximum 1000); unfiltered counts also report the estimated total.

  Paramètres, retour & exemple (5)

  Paramètres

  • name · string · requis
    Collection name.
  • filter · object · optionnel
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • upperBound · integer · optionnel
    Stop counting at this many (Data API maximum 1000).

  Exemple d'appel

  {"name": "count", "arguments": {"name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• distinct_values≈ 369 tokens

  Distinct values of a field (dot paths allowed) across up to scanLimit documents or rows, optionally filtered — useful for discovering categories before filtering.

  Paramètres, retour & exemple (8)

  Paramètres

  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • name · string · requis
    Collection or table name.
  • field · string · requis
    Field path, e.g. "genre" or "author.country".
  • filter · object · optionnel
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • maxValues · integer · optionnel
    Max distinct values returned.
  • scanLimit · integer · optionnel
    Max documents scanned.

  Exemple d'appel

  {"name": "distinct_values", "arguments": {"name": "texte", "field": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• list_vectorize_providers≈ 150 tokens

  List the embedding (vectorize) and reranking providers and models available to the database, with dimensions and authentication options. Use before create_collection with a vectorize service.

  Paramètres, retour & exemple (2)

  Paramètres

  • kind · string · optionnel · embedding | reranking | both
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.

  Exemple d'appel

  {"name": "list_vectorize_providers", "arguments": {"kind": "embedding", "database": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• code_examples≈ 247 tokens

  Canonical Data API client code for Python (astrapy), TypeScript (@datastax/astra-db-ts), Java, C#, and Go from the bundled, documentation-derived example library. Use when writing application code; snippets read ASTRA_DB_APPLICATION_TOKEN / ASTRA_DB_API_ENDPOINT from the environment.

  Paramètres, retour & exemple (4)

  Paramètres

  • mode · string · optionnel · content | list
    content: include source; list: file names only.
  • limit · integer · optionnel
  • query · string · requis
    What the code should do, e.g. "create a vector collection with vectorize", "find with filter and sort".
  • language · string · requis · python | typescript | java | csharp | go
    Client language.

  Exemple d'appel

  {"name": "code_examples", "arguments": {"query": "texte", "language": "python"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• insert≈ 346 tokens

  Insert 1–1000 documents into a collection (or rows into a table), in chunks. On vectorize collections use "$vectorize": "text" to embed server-side; otherwise "$vector": [...] for your own embeddings. Reports inserted ids and any per-document errors.

  Paramètres, retour & exemple (6)

  Paramètres

  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • name · string · requis
    Collection or table name.
  • ordered · boolean · optionnel
    Stop at the first error (default: insert what can be inserted).
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • documents · array · requis
    Documents (collections) or rows (tables). Use "$vectorize": "text" for server-side embeddings on vectorize collections.

  Exemple d'appel

  {"name": "insert", "arguments": {"name": "texte", "documents": []}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• update≈ 461 tokens

  Update the document/row matching a filter with operators ($set, $unset, $inc, $push, …). `many` updates every match (collections only); `many` or an empty filter requires the user's confirmation. `upsert` inserts when nothing matches.

  Paramètres, retour & exemple (9)

  Paramètres

  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • many · boolean · optionnel
    Update every match (collections only). Requires confirmation.
  • name · string · requis
    Collection or table name.
  • filter · object · requis
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • update · object · requis
    Update operators, e.g. {"$set": {"status": "done"}, "$inc": {"views": 1}}.
  • upsert · boolean · optionnel
    Insert when nothing matches (collections).
  • confirm · string · optionnel
    Set ONLY after you described the impact and the user approved it in a later message: the exact target name. The original request is not approval; never guess or self-approve.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.

  Exemple d'appel

  {"name": "update", "arguments": {"name": "texte", "filter": {}, "update": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• delete≈ 351 tokens

  Delete the document/row matching a filter. `many` deletes every match; `many` or an empty filter requires the user's confirmation.

  Paramètres, retour & exemple (7)

  Paramètres

  • kind · string · optionnel · collection | table
    Only needed when a collection and a table share the name; otherwise detected.
  • many · boolean · optionnel
    Delete every match. Requires confirmation.
  • name · string · requis
    Collection or table name.
  • filter · object · requis
    Data API filter, e.g. {"status": "active", "year": {"$gte": 2020}}. Typed ids as {"$uuid": "…"} / {"$objectId": "…"}.
  • confirm · string · optionnel
    Set ONLY after you described the impact and the user approved it in a later message: the exact target name. The original request is not approval; never guess or self-approve.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.

  Exemple d'appel

  {"name": "delete", "arguments": {"name": "texte", "filter": {}}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• create_collection≈ 616 tokens

  Create a collection: plain JSON, vector (dimension + metric), or vectorize (server-side embeddings from a provider — see list_vectorize_providers), optionally with lexical (BM25) and rerank for hybrid search, indexing allow/deny, and a default id type. Idempotent when the settings match an existing collection.

  Paramètres, retour & exemple (8)

  Paramètres

  • name · string · requis
    New collection name (letters, digits, underscore).
  • rerank · object · optionnel
    Reranking service (needed for hybrid).
  • vector · object · optionnel
    Omit for a non-vector collection.
  • lexical · boolean · optionnel
    Enable BM25 lexical search (needed for hybrid).
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • indexing · object · optionnel
    Index only (allow) or all but (deny) these fields. Deny large text fields you never filter on.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • defaultId · string · optionnel · objectId | uuid | uuidv6 | uuidv7
    Auto-generated _id type.

  Exemple d'appel

  {"name": "create_collection", "arguments": {"name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• create_table≈ 436 tokens

  Create a table with typed columns (text, int, uuid, timestamp, map/list/set, vector…) and a primary key (partition columns plus optional clustering order). Then add indexes with create_index for columns you filter or vector-search on.

  Paramètres, retour & exemple (6)

  Paramètres

  • name · string · requis
    New table name.
  • columns · object · requis
    Column types, e.g. {"id": "uuid", "title": "text", "tags": {"type": "set", "valueType": "text"}, "embedding": {"type": "vector", "dimension": 1024}}.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • primaryKey · ? · requis
    A column name, or {"partitionBy": ["user_id"], "partitionSort": {"created_at": -1}}.
  • ifNotExists · boolean · optionnel

  Exemple d'appel

  {"name": "create_table", "arguments": {"name": "texte", "columns": {}, "primaryKey": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• create_index≈ 348 tokens

  Create an index on a table column: regular (filtering and sorting), vector (similarity search on a vector column), or text (BM25 lexical search). Collections index fields automatically.

  Paramètres, retour & exemple (8)

  Paramètres

  • name · string · requis
    Index name.
  • type · string · optionnel · regular | vector | text
    regular (filtering), vector (similarity), text (BM25).
  • table · string · requis
    Table name (collections index automatically; use indexing on create_collection).
  • column · string · requis
    Column to index.
  • options · object · optionnel
    e.g. {"metric": "dot_product"} for vector, {"caseSensitive": false} for regular.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.
  • ifNotExists · boolean · optionnel

  Exemple d'appel

  {"name": "create_index", "arguments": {"name": "texte", "table": "texte", "column": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

• drop≈ 252 tokens

  Permanently drop a collection, a table, or a table index, with all its data. Always requires the user's explicit confirmation.

  Paramètres, retour & exemple (5)

  Paramètres

  • kind · string · requis · collection | table | index
    What to drop.
  • name · string · requis
    Name of the collection, table, or table index.
  • confirm · string · optionnel
    Set ONLY after you described the impact and the user approved it in a later message: the exact target name. The original request is not approval; never guess or self-approve.
  • database · string · optionnel
    Database to use: name, id, or Data API endpoint URL. Omit for the configured default.
  • keyspace · string · optionnel
    Keyspace. Omit for the configured/default keyspace.

  Exemple d'appel

  {"name": "drop", "arguments": {"kind": "collection", "name": "texte"}}

  Exemple de retour

  {"content": [{"type": "text", "text": "…"}]}

Badge pour votre README

Ce serveur a passé le pentest comportemental. Le badge porte son nom et renvoie à cette fiche — la preuve, c'est la fiche (verdict live qu'on contrôle), pas l'image.

[![Vérifié par MCP Hub](https://dokore.fr/badge/io.github.erichare--astra-mcp.svg)](https://dokore.fr/serveur/io.github.erichare--astra-mcp)

<a href="https://dokore.fr/serveur/io.github.erichare--astra-mcp"><img src="https://dokore.fr/badge/io.github.erichare--astra-mcp.svg" alt="Vérifié par MCP Hub" height="20"></a>

Réévalué en continu : si le comportement change, le verdict tombe et le badge disparaît. Un badge qui ne pointe pas vers une fiche dokore.fr ne prouve rien. Comment ça marche

C'est votre serveur ?

Revendiquez-le sur l'espace créateur (preuve GitHub, domaine ou npm) pour afficher le badge « propriétaire vérifié » et lancer un pentest comportemental à la demande afin d'obtenir le badge « Vérifié ».

Créer un compte / revendiquer →

Source

@erichare/astra-mcp@2.0.0