Skip to main content
Glama

Brapi Find Studies

brapi_find_studies
Read-only

Locate studies matching crop, trial type, season, location, program, or recorded observation variable. Enriches results with program/trial/location context in one call. results[] holds the leading matches up to loadLimit and a 24,000-byte budget; when it holds fewer rows than matched, the matched set is materialized as a dataframe — read it with brapi_dataframe_describe and brapi_dataframe_query (SQL).

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "brapi_find_studies"
  }
}

Input Schema

NameRequiredDescriptionDefault
cropNoFilter by common crop name (single value).
aliasNoConnection alias. Omit to use alias `default`. An alias registered via brapi_connect works, and so does one the server has a URL configured for (a built-in alias, an operator-configured alias, or `default` on a deployment with a default server), which connects on first use. Calls that used a non-default alias must pass that same alias here.
activeNoRestrict to active / inactive studies.
trialsNoFilter by trialDbIds.
seasonsNoFilter by seasons (e.g. "2022").
programsNoFilter by programDbIds.
loadLimitNoCap on rows returned inline in results[]. Omit for the deployment default. results[] also stops at a 24,000-byte budget, whichever bound comes first, and keeps at least one row when any matched. When results[] holds fewer rows than matched, the matched rows are staged in a dataframe (marked truncated if the upstream walk stopped short); query it with brapi_dataframe_query (SQL) instead of paging row-by-row.
locationsNoFilter by locationDbIds (server-side identifiers, not display names). On servers that ignore this filter on /studies, it is applied client-side over the scanned studies, and a study with no locationDbId never matches.
variablesNoFilter by observationVariableDbIds — studies that recorded these variables. Get IDs from brapi_find_variables.
studyNamesNoFilter by study display name.
trialTypesNoFilter by study types.
extraFiltersNoExtra BrAPI filters forwarded verbatim. Valid keys vary by endpoint; brapi_describe_filters enumerates them. Named params on this tool take precedence on conflict.

Output Schema

NameRequiredDescriptionDefault
capNoThe effective loadLimit when truncated; shown is below it when the 24,000-byte inline budget bound first (notice says which).
aliasNoConnection alias the call used.
errorNoPresent when the call failed. Absent on success.
shownNoLength of results[] when truncated.
noticeNoWhy results[] is empty, or which bound cut it short and the dataframe holding the matched rows.
hasMoreNoTrue when more rows exist beyond the returned set.
resultsNoLeading matched rows, up to loadLimit and a 24,000-byte budget, whichever binds first.
warningsNoAdvisory messages (filter overrides, partial data).
dataframeNoThe matched rows, staged as a dataframe when results[] holds fewer.
truncatedNoTrue when results[] holds fewer rows than matched because loadLimit or the inline byte budget bound.
totalCountNoTotal rows reported by the server, else the rows scanned; with `locations` applied client-side, the matched rows fetched.
distributionsNoValue frequency per field across every matched row fetched.
returnedCountNoLength of results[].
appliedFiltersNoFilter map sent upstream (named filters plus extraFilters).
refinementHintNoSuggested refinement when the result set is large.
clientSideFiltersNoFilters this server ignores on /studies, applied to the scanned rows instead of sent upstream, keyed by input name.

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed1 schema field changed
    • addedInput schema / properties / variables
      Added value: +{
      +  "description": "Filter by observationVariableDbIds — studies that recorded these variables. Get IDs from brapi_find_variables.",
      +  "items": {
      +    "pattern": "\\S",
      +    "type": "string"
      +  },
      +  "type": "array"
      +}
  2. Changed2 schema fields changed
    • changedOutput schema / properties / error / properties / data / properties / reason / description
      Previous value: -"Machine-readable failure mode. Declared by this tool: `upstream_not_found`: An uncaught BrAPI data request answered HTTP 404. `upstream_rate_limited`: An uncaught BrAPI data request exhausted retries after HTTP 429. `missing_find_route`: Neither the GET list nor a usable POST search route is advertised. `search_endpoint_disabled`: Only a POST search route marked unusable by the active dialect is advertised. `unknown_alias`: No connection is registered under the requested alias, and the server has no base URL configured for it to connect on first use. `auth_token_exchange_failed`: Connecting the alias on first use, the SGN or OAuth token exchange against the BrAPI /token endpoint failed. `auth_no_access_token`: Connecting the alias on first use, the token endpoint answered without returning an access token. `capability_discovery_unavailable`: Neither /serverinfo nor /calls listed any call, and at least one of them timed out, was unreachable, or answered 429 or a 5xx other than 501, so nothing was cached. `upstream_unauthorized`: The BrAPI server answered HTTP 401: it requires credentials, or rejected the ones sent. `upstream_forbidden`: The BrAPI server answered HTTP 403: the credentials sent, or their absence, lack access. `alias_base_url_invalid`: The base URL configured on the server for this alias is not an absolute http(s) URL, or carries credentials. `private_address_blocked`: The base URL configured for this alias, or its token URL, is on a private or reserved address this server may not reach. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `upstream_not_found`: An uncaught BrAPI data request answered HTTP 404. `upstream_rate_limited`: An uncaught BrAPI data request exhausted retries after HTTP 429. `missing_find_route`: Neither the GET list nor a usable POST search route is advertised. `search_endpoint_disabled`: Only a POST search route marked unusable by the active dialect is advertised. `unknown_alias`: No connection is registered under the requested alias, and the server has no base URL configured for it to connect on first use. `auth_token_exchange_failed`: Connecting the alias on first use, the SGN or OAuth token exchange against the BrAPI /token endpoint failed. `auth_no_access_token`: Connecting the alias on first use, the token endpoint answered without returning an access token. `capability_discovery_not_found`: Both /serverinfo and /calls return HTTP 404. `capability_discovery_unavailable`: Neither /serverinfo nor /calls listed any call, and at least one of them timed out, was unreachable, or answered 429 or a 5xx other than 501, so nothing was cached. `upstream_unauthorized`: The BrAPI server answered HTTP 401: it requires credentials, or rejected the ones sent. `upstream_forbidden`: The BrAPI server answered HTTP 403: the credentials sent, or their absence, lack access. `alias_base_url_invalid`: The base URL configured on the server for this alias is not an absolute http(s) URL, or carries credentials. `private_address_blocked`: The base URL configured for this alias, or its token URL, is on a private or reserved address this server may not reach. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."
    • changedOutput schema / properties / error / properties / data / properties / reason / examples
      Previous value: -[
      -  "upstream_not_found",
      -  "upstream_rate_limited",
      -  "missing_find_route",
      -  "search_endpoint_disabled",
      -  "unknown_alias",
      -  "auth_token_exchange_failed",
      -  "auth_no_access_token",
      -  "capability_discovery_unavailable",
      -  "upstream_unauthorized",
      -  "upstream_forbidden",
      -  "alias_base_url_invalid",
      -  "private_address_blocked",
      -  "all_filters_dropped"
      -]New value: +[
      +  "upstream_not_found",
      +  "upstream_rate_limited",
      +  "missing_find_route",
      +  "search_endpoint_disabled",
      +  "unknown_alias",
      +  "auth_token_exchange_failed",
      +  "auth_no_access_token",
      +  "capability_discovery_not_found",
      +  "capability_discovery_unavailable",
      +  "upstream_unauthorized",
      +  "upstream_forbidden",
      +  "alias_base_url_invalid",
      +  "private_address_blocked",
      +  "all_filters_dropped"
      +]
  3. Changed4 schema fields changed
    • changedOutput schema / properties / dataframe / properties / columnLegend / description
      Previous value: -"Sanitized column name → original upstream key, present only when a column was renamed to be SQL-safe (e.g. `end` → `end_`). SQL uses the sanitized names."New value: +"SQL-safe name → original upstream key for renamed columns. Use the safe names in SQL."
    • changedOutput schema / properties / dataframe / properties / expiresAt / description
      Previous value: -"ISO 8601 time the dataframe metadata is purged; copy rows out before, or re-run the producing tool."New value: +"ISO 8601 expiry of this dataframe’s provenance metadata."
    • changedOutput schema / properties / dataframe / properties / totalCount / description
      Previous value: -"Upstream-reported row total, when known; can exceed rowCount."New value: +"Matched total for complete local scans; else upstream total. Omitted for incomplete local scans or unknown totals."
    • changedOutput schema / properties / dataframe / properties / truncated / description
      Previous value: -"True when the dataframe holds less than the upstream reports: a cap, failed page, deadline, or short upstream ended the pull, or it was skipped. warnings says which. Absent otherwise."New value: +"True for a capped, failed, timed-out, incomplete, or skipped pull; warnings explains why."
  4. Changed2 schema fields changed
    • changedOutput schema / properties / error / properties / data / properties / reason / description
      Previous value: -"Machine-readable failure mode. Declared by this tool: `unknown_alias`: No connection is registered under the requested alias, and the server has no base URL configured for it to connect on first use. `auth_token_exchange_failed`: Connecting the alias on first use, the SGN or OAuth token exchange against the BrAPI /token endpoint failed. `auth_no_access_token`: Connecting the alias on first use, the token endpoint answered without returning an access token. `capability_discovery_unavailable`: Neither /serverinfo nor /calls listed any call, and at least one of them timed out, was unreachable, or answered 429 or a 5xx other than 501, so nothing was cached. `upstream_unauthorized`: The BrAPI server answered HTTP 401: it requires credentials, or rejected the ones sent. `upstream_forbidden`: The BrAPI server answered HTTP 403: the credentials sent, or their absence, lack access. `alias_base_url_invalid`: The base URL configured on the server for this alias is not an absolute http(s) URL, or carries credentials. `private_address_blocked`: The base URL configured for this alias, or its token URL, is on a private or reserved address this server may not reach. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `upstream_not_found`: An uncaught BrAPI data request answered HTTP 404. `upstream_rate_limited`: An uncaught BrAPI data request exhausted retries after HTTP 429. `missing_find_route`: Neither the GET list nor a usable POST search route is advertised. `search_endpoint_disabled`: Only a POST search route marked unusable by the active dialect is advertised. `unknown_alias`: No connection is registered under the requested alias, and the server has no base URL configured for it to connect on first use. `auth_token_exchange_failed`: Connecting the alias on first use, the SGN or OAuth token exchange against the BrAPI /token endpoint failed. `auth_no_access_token`: Connecting the alias on first use, the token endpoint answered without returning an access token. `capability_discovery_unavailable`: Neither /serverinfo nor /calls listed any call, and at least one of them timed out, was unreachable, or answered 429 or a 5xx other than 501, so nothing was cached. `upstream_unauthorized`: The BrAPI server answered HTTP 401: it requires credentials, or rejected the ones sent. `upstream_forbidden`: The BrAPI server answered HTTP 403: the credentials sent, or their absence, lack access. `alias_base_url_invalid`: The base URL configured on the server for this alias is not an absolute http(s) URL, or carries credentials. `private_address_blocked`: The base URL configured for this alias, or its token URL, is on a private or reserved address this server may not reach. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."
    • changedOutput schema / properties / error / properties / data / properties / reason / examples
      Previous value: -[
      -  "unknown_alias",
      -  "auth_token_exchange_failed",
      -  "auth_no_access_token",
      -  "capability_discovery_unavailable",
      -  "upstream_unauthorized",
      -  "upstream_forbidden",
      -  "alias_base_url_invalid",
      -  "private_address_blocked",
      -  "all_filters_dropped"
      -]New value: +[
      +  "upstream_not_found",
      +  "upstream_rate_limited",
      +  "missing_find_route",
      +  "search_endpoint_disabled",
      +  "unknown_alias",
      +  "auth_token_exchange_failed",
      +  "auth_no_access_token",
      +  "capability_discovery_unavailable",
      +  "upstream_unauthorized",
      +  "upstream_forbidden",
      +  "alias_base_url_invalid",
      +  "private_address_blocked",
      +  "all_filters_dropped"
      +]
  5. Changed34 schema fields changed
    • changedInput schema / properties / alias / description
      Previous value: -"Connection alias registered via brapi_connect. Omit to read the connection registered under alias `default` — i.e. a prior brapi_connect call that did not specify an alias. Calls that used a non-default alias must pass that same alias here."New value: +"Connection alias. Omit to use alias `default`. An alias registered via brapi_connect works, and so does one the server has a URL configured for (a built-in alias, an operator-configured alias, or `default` on a deployment with a default server), which connects on first use. Calls that used a non-default alias must pass that same alias here."
    • addedInput schema / properties / crop / pattern
      Added value: +"\\S"
    • changedInput schema / properties / loadLimit / description
      Previous value: -"Cap on rows returned inline. Omit for the deployment default. Rows beyond the cap land in a dataframe; query with brapi_dataframe_query (SQL) instead of paging row-by-row."New value: +"Cap on rows returned inline in results[]. Omit for the deployment default. results[] also stops at a 24,000-byte budget, whichever bound comes first, and keeps at least one row when any matched. When results[] holds fewer rows than matched, the matched rows are staged in a dataframe (marked truncated if the upstream walk stopped short); query it with brapi_dataframe_query (SQL) instead of paging row-by-row."
    • changedInput schema / properties / locations / description
      Previous value: -"Filter by locationDbIds (server-side identifiers, not display names)."New value: +"Filter by locationDbIds (server-side identifiers, not display names). On servers that ignore this filter on /studies, it is applied client-side over the scanned studies, and a study with no locationDbId never matches."
    • addedInput schema / properties / locations / items / pattern
      Added value: +"\\S"
    • addedInput schema / properties / programs / items / pattern
      Added value: +"\\S"
    • addedInput schema / properties / seasons / items / pattern
      Added value: +"\\S"
    • addedInput schema / properties / studyNames / items / pattern
      Added value: +"\\S"
    • addedInput schema / properties / trialTypes / items / pattern
      Added value: +"\\S"
    • addedInput schema / properties / trials / items / pattern
      Added value: +"\\S"
    • changedOutput schema / properties / alias / description
      Previous value: -"Alias of the registered BrAPI connection the call used."New value: +"Connection alias the call used."
    • changedOutput schema / properties / appliedFilters / description
      Previous value: -"The final filter map sent to the server (named + extraFilters)."New value: +"Filter map sent upstream (named filters plus extraFilters)."
    • addedOutput schema / properties / cap
      Added value: +{
      +  "description": "The effective loadLimit when truncated; shown is below it when the 24,000-byte inline budget bound first (notice says which).",
      +  "exclusiveMinimum": 0,
      +  "maximum": 9007199254740991,
      +  "type": "integer"
      +}
    • addedOutput schema / properties / clientSideFilters
      Added value: +{
      +  "additionalProperties": {
      +    "items": {
      +      "type": "string"
      +    },
      +    "type": "array"
      +  },
      +  "description": "Filters this server ignores on /studies, applied to the scanned rows instead of sent upstream, keyed by input name.",
      +  "propertyNames": {
      +    "type": "string"
      +  },
      +  "type": "object"
      +}
    • changedOutput schema / properties / dataframe / description
      Previous value: -"Dataframe handle when the full result set was materialized as a dataframe. Query it with brapi_dataframe_query (SQL)."New value: +"The matched rows, staged as a dataframe when results[] holds fewer."
    • changedOutput schema / properties / dataframe / properties / columnLegend / description
      Previous value: -"Maps a sanitized column name back to its original upstream key, for columns renamed to clear the SQL-safe-identifier gate (e.g. `end` → `end_`). Present only when a column was renamed — write SQL against the sanitized (left-hand) names."New value: +"Sanitized column name → original upstream key, present only when a column was renamed to be SQL-safe (e.g. `end` → `end_`). SQL uses the sanitized names."
    • changedOutput schema / properties / dataframe / properties / columns / description
      Previous value: -"Full column list of the dataframe."New value: +"Every column name."
    • changedOutput schema / properties / dataframe / properties / columns / items / description
      Previous value: -"Column name from the materialized rows."New value: +"Column name."
    • changedOutput schema / properties / dataframe / properties / createdAt / description
      Previous value: -"ISO 8601 timestamp the dataframe was created."New value: +"ISO 8601 creation time."
    • changedOutput schema / properties / dataframe / properties / expiresAt / description
      Previous value: -"ISO 8601 timestamp after which the dataframe metadata will be purged. Re-run the find_* tool to refresh, or copy results out before expiry."New value: +"ISO 8601 time the dataframe metadata is purged; copy rows out before, or re-run the producing tool."
    • changedOutput schema / properties / dataframe / properties / maxRows / description
      Previous value: -"Cap that was applied at create time, when truncation occurred."New value: +"Row cap that ended the pull, present only when that cap fired."
    • changedOutput schema / properties / dataframe / properties / rowCount / description
      Previous value: -"Number of rows materialized in the dataframe."New value: +"Rows in the dataframe."
    • changedOutput schema / properties / dataframe / properties / tableName / description
      Previous value: -"Dataframe name. Use with brapi_dataframe_describe (schema + provenance) and brapi_dataframe_query (SQL)."New value: +"Table name for brapi_dataframe_describe and brapi_dataframe_query."
    • changedOutput schema / properties / dataframe / properties / totalCount / description
      Previous value: -"Total rows reported by the upstream server (may exceed rowCount when truncated)."New value: +"Upstream-reported row total, when known; can exceed rowCount."
    • changedOutput schema / properties / dataframe / properties / truncated / description
      Previous value: -"True when the dataframe hit a row cap before exhausting upstream."New value: +"True when the dataframe holds less than the upstream reports: a cap, failed page, deadline, or short upstream ended the pull, or it was skipped. warnings says which. Absent otherwise."
    • changedOutput schema / properties / distributions / description
      Previous value: -"Value frequency per field across the full result set."New value: +"Value frequency per field across every matched row fetched."
    • changedOutput schema / properties / error / properties / data / properties / reason / description
      Previous value: -"Machine-readable failure mode. Declared by this tool: `unknown_alias`: No connection has been registered under the requested alias. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."New value: +"Machine-readable failure mode. Declared by this tool: `unknown_alias`: No connection is registered under the requested alias, and the server has no base URL configured for it to connect on first use. `auth_token_exchange_failed`: Connecting the alias on first use, the SGN or OAuth token exchange against the BrAPI /token endpoint failed. `auth_no_access_token`: Connecting the alias on first use, the token endpoint answered without returning an access token. `capability_discovery_unavailable`: Neither /serverinfo nor /calls listed any call, and at least one of them timed out, was unreachable, or answered 429 or a 5xx other than 501, so nothing was cached. `upstream_unauthorized`: The BrAPI server answered HTTP 401: it requires credentials, or rejected the ones sent. `upstream_forbidden`: The BrAPI server answered HTTP 403: the credentials sent, or their absence, lack access. `alias_base_url_invalid`: The base URL configured on the server for this alias is not an absolute http(s) URL, or carries credentials. `private_address_blocked`: The base URL configured for this alias, or its token URL, is on a private or reserved address this server may not reach. `all_filters_dropped`: The active dialect dropped every filter the agent supplied — the upstream server does not honor any of the requested scope filters on this endpoint, so the call would silently widen to the unfiltered baseline. Other values are possible when a failure originates below the handler."
    • changedOutput schema / properties / error / properties / data / properties / reason / examples
      Previous value: -[
      -  "unknown_alias",
      -  "all_filters_dropped"
      -]New value: +[
      +  "unknown_alias",
      +  "auth_token_exchange_failed",
      +  "auth_no_access_token",
      +  "capability_discovery_unavailable",
      +  "upstream_unauthorized",
      +  "upstream_forbidden",
      +  "alias_base_url_invalid",
      +  "private_address_blocked",
      +  "all_filters_dropped"
      +]
    • changedOutput schema / properties / notice / description
      Previous value: -"Guidance when no rows were returned — how to broaden filters or retry."New value: +"Why results[] is empty, or which bound cut it short and the dataframe holding the matched rows."
    • changedOutput schema / properties / refinementHint / description
      Previous value: -"Suggested next-step query refinement when the result set is large."New value: +"Suggested refinement when the result set is large."
    • changedOutput schema / properties / results / description
      Previous value: -"Rows returned in-context (up to loadLimit)."New value: +"Leading matched rows, up to loadLimit and a 24,000-byte budget, whichever binds first."
    • addedOutput schema / properties / shown
      Added value: +{
      +  "description": "Length of results[] when truncated.",
      +  "exclusiveMinimum": 0,
      +  "maximum": 9007199254740991,
      +  "type": "integer"
      +}
    • changedOutput schema / properties / totalCount / description
      Previous value: -"Total rows reported by the server."New value: +"Total rows reported by the server, else the rows scanned; with `locations` applied client-side, the matched rows fetched."
    • addedOutput schema / properties / truncated
      Added value: +{
      +  "description": "True when results[] holds fewer rows than matched because loadLimit or the inline byte budget bound.",
      +  "type": "boolean"
      +}
  6. First observed

TDQS

A3.8/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations only declare readOnlyHint and openWorldHint; the description adds real behavioral context beyond that — enrichment with program/trial/location data in one call, a 24,000-byte inline budget, a matched-set dataframe fallback, and a truncated marker when the upstream walk stops short. It omits any note on permissions, rate limits, or server variability beyond the locations caveat carried in the schema.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness5/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Three sentences, front-loaded with the purpose, then the enrichment value, then the results[]/dataframe contract. The loadLimit budget detail is duplicated in the schema, but it is essential enough to justify being surfaced up front.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 12-parameter, zero-required, open-world search tool with an output schema, the definition covers what an agent most needs: filter surface, inline result budget, and the escape hatch to the dataframe tools. A brief note on the single-study alternative would make it complete.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and each parameter already carries a descriptive string, including the alias, locations, and loadLimit caveats. The description adds only aggregate framing ('matching crop, trial type, season...') and does not deepen individual parameter meaning beyond the schema, so the baseline 3 applies.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

Names a specific verb (locate) and resource (studies) and enumerates the filterable dimensions (crop, trial type, season, location, program, observation variable), which lets an agent match intent quickly. It stops short of distinguishing itself from siblings like brapi_get_study or brapi_federated_find, so the boundary is inferred rather than stated.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives clear usage guidance for the truncation case: when results[] holds fewer rows than matched, query the staged dataframe with brapi_dataframe_query instead of paging row-by-row. However, it never says when to prefer this over brapi_get_study or brapi_federated_find, and there are no explicit exclusions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.