Salesforce Safe Delivery

    1

    Free guard rails for AI on live Salesforce orgs, from development to deployment: standard-first design, sandbox builds, tested Apex, LWC and Flow, impact analysis and safe releases.

    Free

    Security scanned

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIMuseMuseDotsDots+21 more

    Salesforce Safe Delivery

    Example session with this skill installed

    I need to add a new validation rule to the Account object to ensure the Tax ID field is populated when the Type is 'Prospect'. We use a GitHub Actions pipeline for deployments.

    • Read your context and instructions
    • Compiled the salesforce safe delivery

    I have analyzed the requirement and the current environment. Since you are using a GitHub Actions pipeline, I will proceed in

    Mode A (Pipeline/CAB): I prepare the change, your pipeline deploys it. I have verified the live Account metadata in production (read-only) and in the sandbox.

    Proposed Change: Account Validation Rule

    OptionSolutionTrade-off
    Standard ConfigValidation RuleRecommended. Simple to maintain and stops bad data at the source.
    DeclarativeScreen FlowOverkill for a single required-field rule.
    Custom CodeApex TriggerNot recommended. Adds technical debt for a standard validation.

    Impact Analysis

    • Metadata: adds Account.Tax_ID_Required_for_Prospect.validationRule-meta.xml.

    Existing data: SELECT COUNT() FROM Account WHERE Type = 'Prospect' AND Tax_ID__c = null returns

    42 in production. Any update to those records will fail until the Tax ID is filled in. Only the count was read; no record values entered this session.

    Lead conversion: if "Require Validation for Converted Leads" is enabled, conversions to Prospect accounts without a Tax ID will fail.

    • Integrations: an ERP or other sync that updates Prospect accounts without a Tax ID will start getting errors.

    Implementation Status

    Discovery: checked existing ValidationRule metadata on Account in production and sandbox. No conflicting rules.
    2. Drafted rule: AND( ISPICKVAL(Type, 'Prospect'), ISBLANK(Tax_ID__c) )
    3.

    Rollback: deactivate the rule through the pipeline as an emergency change. No data is changed, so nothing else needs undoing.

    Next steps

    1. Decide how to handle the 42 existing Prospect accounts: clean them up first, or accept errors on their next update.
    2. I'll deploy the rule to dev-sandbox for functional testing, including a lead conversion and an integration update.
    3. Then I'll prepare the Pull Request and the CAB release note. Your pipeline deploys to production after approval.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    Guides your AI agent through the whole Salesforce delivery: understanding the request, designing with standard features first, building and testing in a sandbox, and releasing safely. It stops the AI from writing unnecessary custom code, deploying over another team's change, firing automation nobody mapped, or touching production without approval.

    • Develops like a careful architect: live-org discovery, an options table (standard → Flow → product → custom), a short design spec, small tested tasks and end-to-end UI testing.
    • Fits your release process. In teams with a CI/CD pipeline and change board, the AI prepares the PR, impact analysis and CAB release note, and never deploys to production. Teams without a pipeline can let it deploy after an explicit GO.
    • Production read-only; no personal data pulled into the AI session.
    • Verifies the live org, not just git. Protects shared components.
    • Impact analysis for triggers, flows, validation rules and email cascades.
    • Release gates: validation, rollback plan, data snapshot.

    Free and open source (MIT): https://github.com/pradeep-chandramohan/salesforce-safe-delivery

    Requires Salesforce CLI (sf) and active org auth. AI cannot bypass manual Setup steps. Production DML requires explicit user GO. Sensitive data masking is the user's responsibility.

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before any payment.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Free forever

    No account required to browse

    Trust & safety

    Security scan

    Security scan passed today

    • Free with an Agensi account

    Requires

    Salesforce
    Git
    salesforce CLI

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Network Access

    Allowed Hosts

    login.salesforce.com

    Runs Salesforce CLI (sf) and git commands in your Salesforce project. Connects only to the orgs you have authorized. Production is read-only by default; deploys need your explicit approval or go through your pipeline.

    Listedtoday

    What's inside

    Frequently Asked Questions

    This skill acts as a safety layer for Claude Code, enforcing mandatory sandbox validation and impact analysis to prevent the AI from making direct, unverified changes that could break production metadata or automation.

    It is designed to work with Claude Code via the Salesforce CLI (sf commands) and integrates seamlessly with popular DevOps tools like Gearset, Copado, and GitHub Actions.

    You will receive the skill configuration file containing the specialized safety protocols, deployment checklists, and environment discovery logic ready to be imported into your AI agent environment. It's free.

    While the skill handles the release logic, you must provide your own Salesforce CLI authentication and have the necessary permissions to access the target sandboxes and production orgs.

    Yes, it is designed to recognize and adapt to your existing CI/CD setup, functioning as a 'Preparer' for enterprise Change Advisory Board (CAB) workflows or an 'Operator' for smaller org deployments.

    The skill performs 'live discovery' via the Tooling API to verify the current state of your org's metadata, ensuring it doesn't rely on outdated git history before suggesting changes.

    New in DevOps & Cloud