Using Jag's SF Skills

Jag Valaiyapathy, a Salesforce Certified Technical Architect (CTA), has built and open-sourced the most comprehensive Salesforce skill library available. It covers 33 skills across Apex, Flow, LWC, SOQL, metadata, Data 360, Agentforce, and OmniStudio, each with built-in validation rubrics and scoring.

What's Included

The library organizes skills by capability area:

Development

sf-apex · sf-flow · sf-lwc · sf-soql: Core development skills with scoring rubrics for Apex (90-pt), Flow (110-pt), LWC (165-pt), and SOQL (100-pt)

Quality

sf-testing · sf-debug: Test execution, coverage analysis, bulk testing patterns, debug log analysis, and governor limit troubleshooting

Foundation

sf-metadata · sf-data · sf-docs · sf-permissions: Metadata generation, data operations, official docs retrieval, and permission set analysis

Integration

sf-connected-apps · sf-integration: OAuth flows, External Client Apps, Named Credentials, REST/SOAP callouts, Platform Events, and Change Data Capture

AI & Automation

sf-ai-agentscript · sf-ai-agentforce · sf-ai-agentforce-testing · sf-ai-agentforce-observability · sf-ai-agentforce-persona: Agent design, Agent Script DSL, testing, trace analysis, and persona design

Data 360

sf-datacloud + 6 phase-specific skills. End-to-end Data 360 pipeline: connect, prepare, harmonize, segment, act, and retrieve

Industries

sf-industry-commoncore-omnistudio-analyze + 5 OmniStudio skills. Data Mappers, Integration Procedures, OmniScripts, FlexCards, and callable Apex

DevOps

sf-deploy · sf-diagram-mermaid · sf-diagram-nanobananapro: Deployment automation, Mermaid diagrams, and visual artifacts

Installation

npx skills add Jaganpro/sf-skills

curl -sSL https://raw.githubusercontent.com/Jaganpro/sf-skills/main/tools/install.sh | bash

npx skills add Jaganpro/sf-skills --skill sf-apex

npx skills add Jaganpro/sf-skills --skill sf-apex
npx skills add Jaganpro/sf-skills --skill sf-lwc
npx skills add Jaganpro/sf-skills --skill sf-flow

The FDE pod (Claude Code subagents)

The full installer ships a pre-configured team of 7 Claude Code subagents under ~/.claude/agents/. It's modeled after a real Forward Deployed Engineering pod: one strategist who plans and delegates, two FDE specialists who build Agentforce-specific work, two cross-cutting specialists who handle QA and release for everyone, and two PS (Platform Services) architects who own the platform infrastructure underneath.

fde-strategist                     Plans, delegates up to 4 concurrent workers, no edits
├── FDE team (agent-specific)
│   ├── fde-engineer               Bot metadata, topics, actions, Apex, Agent Scripts
│   └── fde-experience-specialist  Persona, conversation design, utterances, LWC
├── Cross-cutting (both teams)
│   ├── fde-qa-engineer            Apex + agent testing, debug logs, session tracing
│   └── fde-release-engineer       Deploys, Connected Apps, Agent Script CLI, CI/CD
└── PS team (platform infrastructure)
    ├── ps-technical-architect     Service Apex, integrations, data, LWC, performance
    └── ps-solution-architect      Metadata, Flows, permissions, Mermaid diagrams

Each subagent is a single .md file with YAML frontmatter declaring its model, permissionMode, tools, disallowedTools, and bound skills. The strategist runs in plan mode with Edit and Write disabled. The six implementers run in acceptEdits with a constrained toolset.

Who does what

Using the pod

You don't invoke the specialists directly. You ask the strategist, and it plans and delegates:

Use the fde-strategist agent. Goal: build a Customer Service Triage
Agentforce agent for ACME in the acme-uat org. Topics for warranty,
billing, and shipping. Include a companion LWC, a Connected App for the
mobile channel, and an agent test suite covering all three topics plus
an ambiguous utterance. Target org alias: acme-uat.

The strategist returns a plan and then dispatches: ps-solution-architect drafts metadata and the Mermaid diagram, fde-engineer scaffolds the bot, fde-experience-specialist writes the persona and LWC in parallel, ps-technical-architect builds service Apex, fde-qa-engineer creates tests, fde-release-engineer runs a dry-run deploy. Up to four of those run concurrently.

The full playbook on when to use subagents and how to write your own briefs lives in Subagents & Parallelism. That page also covers how to reproduce the pod's behavior in Cursor with hand-written briefs that reference the same skills.

Prerequisites

The skills themselves only require Node.js 18+ to install. However, certain features need additional tools:

How the Skills Work in Practice

Once installed, the skills activate automatically based on what you're working on. Here's what changes in your daily workflow:

Apex Development

Without skills, you ask Cursor to write a trigger and get something functional but generic.

With sf-apex loaded, Cursor generates code that follows:

• Trigger Actions Framework patterns with proper handler separation
• Bulkification with collection-based processing
• Custom exception classes and structured error handling
• Test classes with bulk scenarios (200+ records) and meaningful assertions

The skill scores output on a 90-point rubric across 5 categories: Testing (25), Architecture (20), Clean Code (20), Error Handling (15), and Performance (10).

LWC Development

With sf-lwc loaded, Cursor generates components that follow:

• SLDS 2 class usage and styling hooks
• Accessibility patterns (ARIA labels, keyboard navigation, focus management)
• Dark mode compatibility
• Wire service best practices
• Jest test scaffolding

The skill scores on a 165-point rubric across 9 categories, covering everything from component structure to GraphQL patterns.

Flow Development

With sf-flow loaded, Cursor builds Flows with:

• Proper fault paths on every DML and callout element
• Naming conventions for variables and decision elements
• Bulkification considerations
• Entry condition best practices

Scored on a 110-point rubric covering Design/Naming, Logic/Structure, Error Handling, Architecture, Security, and Performance.

Usage Examples

Here are real prompts you can use once the skills are installed. The AI will automatically apply the relevant skill's guidance.

Apex

/sf-apex Generate an Apex trigger for Account using Trigger Actions Framework

/sf-apex Review my AccountService class for best practices

/sf-apex Create a batch job to process millions of records with proper error handling

LWC

/sf-lwc Create a datatable component to display Accounts with sorting and pagination

/sf-lwc Build a form component for creating Contacts with field validation

/sf-lwc Generate a Jest test for my accountCard component

Flow

/sf-flow Create a record-triggered flow for opportunity stage changes with error handling

/sf-flow Build a screen flow for account creation with validation

SOQL

/sf-soql Optimize this query: SELECT Id, Name FROM Account WHERE Name LIKE '%Corp%'

/sf-soql Generate a SOQL query to find duplicate Leads by Email

Agentforce

/sf-ai-agentforce Create an Agentforce agent for customer support triage

/sf-ai-agentforce Generate a PromptTemplate for case summaries

Integration

/sf-connected-apps Create a Named Credential for Stripe API with OAuth client credentials

/sf-integration Generate a REST callout service with retry and error handling

Updating

npx skills check
npx skills update

python3 ~/.claude/sf-skills-install.py --status
python3 ~/.claude/sf-skills-install.py --update

python3 ~/.claude/sf-skills-install.py --force-update

Video Tutorials

Jag has published video walkthroughs of the skills in action:

Full Skill Reference

Here's the complete list of every skill in the library with its current status:

Development & Quality

Foundation & Integration

AI & Automation

Data 360

Industries (OmniStudio)

DevOps & Diagrams

Contributing & Support

The library is fully open source under the MIT License. If you find an issue or want to contribute:

• GitHub: github.com/Jaganpro/sf-skills
• Issues: github.com/Jaganpro/sf-skills/issues
• Contributing guide: CONTRIBUTING.md