AwesomeSalesforceSkills

Overview

SfSkills is a universal knowledge layer specifically designed for Salesforce AI coding assistants. It consists of three layers: 978+ structured skills with source-grounded instructions, shared templates and canonical building blocks, and a live-org MCP server with 38 tools. The project enables AI assistants to provide role-accurate, Salesforce-specific guidance across all major roles (Admin, Developer, BA, Data, Architect) and Salesforce clouds. The MCP server extends functionality with live-org metadata retrieval, SOQL query capabilities, and diagnostic tools.

Try asking AI

After installing, here are 5 things you can ask your AI assistant:

• skill_search

  Search for Salesforce skills in the knowledge library

  Try:Find skills for trigger frameworksTry:Search for permission set patterns
• agent_suggest

  Suggest appropriate agent based on current task

  Try:Which agent should handle user access diff?Try:Suggest an agent for apex debugging
• metadata_retrieve

  Retrieve metadata from live Salesforce org

  Try:Get custom fields from Account objectTry:Retrieve validation rules
• soql_query

  Execute SOQL queries against live Salesforce org

  Try:Query top 10 accounts by revenueTry:Find all active users created this month
• health_check

  Check health status of MCP server

  Try:Check server healthTry:Get system diagnostics
• apex_ref_probe

  Probe for Apex class and trigger references

  Try:Find references to AccountHandler classTry:Check trigger recursion patterns
• flow_ref_probe

  Probe for Flow references in the org

  Try:Find all flows that update opportunitiesTry:Check flow fault paths
• permset_shape_probe

  Probe permission set shapes and permissions

  Try:Analyze admin permission setTry:Check permission set group assignments
• automation_graph

  Retrieve automation graph and dependencies

  Try:Show automation dependencies for lead processingTry:Map approval process flows
• template_retrieve

  Retrieve reusable Salesforce templates

  Try:Get trigger framework templateTry:Retrieve LWC component skeleton
• decision_tree_lookup

  Lookup decision trees for automation and sharing

  Try:Get automation decision treeTry:Find sharing rules decision path
• envelope_persist

  Persist data envelopes for atomic operations

  Try:Persist this configuration envelopeTry:Save this metadata package

Comparable tools

Installation

# 1. Clone the repository
git clone https://github.com/PranavNagrecha/AwesomeSalesforceSkills.git
cd AwesomeSalesforceSkills

# 2. Install dependencies
python3 -m pip install -r requirements.txt

# 3. Install MCP server
pip install sfskills-mcp

# For Claude Desktop:
# Add to claude_desktop_config.json:
{
  "mcpServers": {
    "sfskills": {
      "command": "sfskills-mcp",
      "args": []
    }
  }
}

FAQ

What AI tools are supported?

SfSkills supports Claude Code, Cursor, Windsurf, Augment, and Codex CLI directly, with Aider and any LLM via direct skill access.

How does the live-org MCP server work?

The MCP server provides 38 tools for searching skills/agents/templates, retrieving live-org metadata, executing SOQL queries, and diagnosing org health, using your authenticated Salesforce CLI connection.

Compare AwesomeSalesforceSkills with