io.github.conquext/neuron

What we checked

Six static checks, weighted by risk. Every result reflects only what could be observed in the published package and repository — never intent.

• Injection surface

  fail25% of grade

  Tool descriptions/manifest scanned for instruction-injection patterns (imperatives at the model, hidden text, 'ignore previous', data-exfil URLs).

  tool "neuron_update_webhook" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_delete_webhook" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_get_webhook_logs" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_get_webhook_logs" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_update_outbound_webhook" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_delete_outbound_webhook" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_get_outbound_webhook_logs" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_get_outbound_webhook_logs" param "id" description: names an outbound copy/exfil destination (BCC/exfil verb). tool "neuron_switch_instance" param "url" description: off-vendor URL in a description (api.client.com); capability text should not name where data goes. tool "neuron_publish_to_pool" param "defaultSecrets" description: off-vendor URL in a description (...); capability text should not name where data goes. tool "neuron_update_pool_item" param "defaultSecrets" description: off-vendor URL in a description (...); capability text should not name where data goes.

• Supply chain

  pass25% of grade

  Package provenance: namespace verification, repo linkage, maintainer count, account age, postinstall scripts, typosquat distance.

  repository linkage verified

• Credential hygiene

  fail15% of grade

  How the server takes secrets (env vs plaintext config vs hardcoded); secrets appearing in tool schemas.

  tool "neuron_register" has a secret-shaped input parameter "password" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_refresh_token" has a secret-shaped input parameter "refreshToken" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_bot_api_chat" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_bot_api_send" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_bot_api_list_conversations" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_bot_api_get_messages" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_bot_api_get_bot" has a secret-shaped input parameter "apiKey" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_create_channel" has a secret-shaped input parameter "metaAccessToken" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_update_channel" has a secret-shaped input parameter "metaAccessToken" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_publish_to_pool" has a secret-shaped input parameter "defaultSecrets" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_update_pool_item" has a secret-shaped input parameter "defaultSecrets" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_update_whatsapp_credentials" has a secret-shaped input parameter "whatsappAccessToken" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_create_social_channel" has a secret-shaped input parameter "credentials" — secrets must never be tool parameters (the model would supply/handle the credential). tool "neuron_update_social_channel" has a secret-shaped input parameter "credentials" — secrets must never be tool parameters (the model would supply/handle the credential).

• Permission scope

  pass15% of grade

  Declared tools vs. breadth (filesystem, network, exec); flags shell-exec and unbounded filesystem access.

  Network egress present (package.json depends on HTTP client "axios") — expected for most servers and not a finding on its own. No shell-exec, unbounded/sensitive filesystem, or download-and-run primitives found in scanned source. PASS = no static red flag. Static scanning is blind to code fetched or executed at runtime.

• Version behavior

  unscannable10% of grade

  Diff of tool definitions between versions; new permissions or changed descriptions in a patch release (the postmark-mcp class).

  no prior version to diff (first sight of io.github.conquext/[email protected]); version-drift / rug-pull cannot be evaluated. Cold-start risk is covered by point-in-time checks (injection_surface, supply_chain, credential_hygiene), not here.

• Transport config

  fail10% of grade

  Remote servers: TLS and auth mode (none/token/OAuth). Local servers: whether the manifest indicates it phones home.

  remote MCP endpoint declares NO authentication (unauthenticated remote — nginx-ui/CVE-2026-33032 class)