Salesforce ConnectorSAFE
MCP Salesforce connector
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
A Model Context Protocol (MCP) server implementation for Salesforce integration, allowing LLMs to interact with Salesforce data through SOQL queries and SOSL searches.
Features
- Execute SOQL (Salesforce Object Query Language) queries
- Perform SOSL (Salesforce Object Search Language) searches
- Retrieve metadata for Salesforce objects, including field names, labels, and types
- List all available SObjects - Discover standard and custom objects
- Retrieve, create, update, and delete records
- Execute Tooling API requests
- Execute Apex REST requests
- Make direct REST API calls to Salesforce
Configuration
Model Context Protocol
To use this server with the Model Context Protocol, you need to configure it in your claude_desktop_config.json file. Add the following entry to the mcpServers section:
{ "mcpServers": { "salesforce": { "command": "uvx", "args": [ "--from", "mcp-salesforce-connector", "salesforce" ], "env": { "SALESFORCEACCESSTOKEN": "SALESFORCEACCESSTOKEN", "SALESFORCEINSTANCEURL": "SALESFORCEINSTANCEURL", "SALESFORCEDOMAIN": "SALESFORCEDOMAIN" } } } }
Available Tools
Query and Search Tools
- `run_soql_query` - Execute SOQL queries against Salesforce
- `run_sosl_search` - Perform SOSL searches across objects
Metadata Tools
- `get_object_fields` - Retrieve field metadata for specific objects
- `list_sobjects` - List all available SObjects (standard and custom)
Single Record Operations
- `get_record` - Retrieve a specific record by ID
- `create_record` - Create a new record
- `update_record` - Update an existing record
- `delete_record` - Delete a record
Bulk Operations
- `bulk_create_records` - Create multiple records in a single o
e495a89f256cOBSERVED · 2026-10-06Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add mcp-salesforce-connector --env SALESFORCE_ACCESS_TOKEN=${SALESFORCE_ACCESS_TOKEN} --env SALESFORCE_CLIENT_SECRET=${SALESFORCE_CLIENT_SECRET} --env SALESFORCE_PASSWORD=${SALESFORCE_PASSWORD} --env SALESFORCE_SECURITY_TOKEN=${SALESFORCE_SECURITY_TOKEN} -- uvx mcp-salesforce-connector{
"mcpServers": {
"mcp-salesforce-connector": {
"command": "uvx",
"args": [
"mcp-salesforce-connector"
],
"env": {
"SALESFORCE_ACCESS_TOKEN": "${SALESFORCE_ACCESS_TOKEN}",
"SALESFORCE_CLIENT_SECRET": "${SALESFORCE_CLIENT_SECRET}",
"SALESFORCE_PASSWORD": "${SALESFORCE_PASSWORD}",
"SALESFORCE_SECURITY_TOKEN": "${SALESFORCE_SECURITY_TOKEN}"
}
}
}
}Exposed tools (14)
4 read · 8 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
apex_execute | write | Executes an Apex REST request |
bulk_create_records | write | Creates multiple records of a specified SObject type in bulk. |
bulk_delete_records | destructive | Deletes multiple records of a specified SObject type in bulk, given their IDs. |
bulk_update_records | write | Updates multiple records of a specified SObject type in bulk. Each record must have an |
create_record | write | Creates a new record |
delete_record | destructive | Deletes a record |
get_object_fields | read | |
get_record | read | |
list_sobjects | read | Retrieves a list of all available Salesforce SObjects (standard and custom). |
restful | read | Makes a direct REST API call to Salesforce |
run_soql_query | write | |
run_sosl_search | write | |
tooling_execute | write | Executes a Tooling API request |
update_record | write | Updates an existing record |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (1 observation(s))
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- pinned
- Secrets in source
- none-found
Findings (1)
bulk_delete_records, delete_record
Gates applied: no_behavioural_pass.
e495a89f256cfull audit observations/trust-audit/mcp-server/smn2gnt__salesforce-connector.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-06 | e495a89f256c | SAFE | B | 89 | first audit |
Questions
What is the Salesforce Connector MCP server?
MCP Salesforce connector
What tools does Salesforce Connector expose?
14 in total: 4 read-only, 8 that write, and 2 that can delete or overwrite (bulk_delete_records, delete_record). Every one is listed on this page with its risk.
Is Salesforce Connector safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Salesforce Connector need?
It reads SALESFORCE_ACCESS_TOKEN, SALESFORCE_CLIENT_SECRET, SALESFORCE_PASSWORD and SALESFORCE_SECURITY_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Salesforce Connector run?
It speaks stdio, so it runs as a local process your client starts. It is published on PyPI as mcp-salesforce-connector.
How current is this page?
The grade is for one exact copy of the source (e495a89f256c), read on 2026-10-06. The repository is watched and re-audited when it changes.