Platform Sandbox Configure logo

Platform Sandbox Configure

OrganizationPopular
forcedotcom
platform-sandbox-configure

MUST USE this skill for ANY sandbox request — including simply getting a sandbox's details, status, license type, or pending-activation state by name or ID. TRIGGER when the user: types "sandbox -help"/"sandbox help"; mentions a sandbox ID (07E prefix); asks to list or show sandboxes; asks for the details, status, license type, or config of a sandbox by name or ID; activates or discards a completed refresh; deletes a sandbox; verifies activation or deletion; creates or refreshes a sandbox. DO NOT TRIGGER when: the user wants to clone a sandbox.

Overview

Publisherforcedotcom
Repositorysf-skills
Skill nameplatform-sandbox-configure
Stars
1.1K
Forks
351
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by forcedotcom on GitHub. Read the source before you install it.

Installation

Install the Platform Sandbox Configure AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/forcedotcom/sf-skills/tree/main/plugins/builder/dx-org-lifecycle/skills/platform-sandbox-configure
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/forcedotcom/sf-skills.git /tmp/sf-skills
mkdir -p .claude/skills
cp -r /tmp/sf-skills/plugins/builder/dx-org-lifecycle/skills/platform-sandbox-configure .claude/skills/platform-sandbox-configure
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Platform Sandbox Configure in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Platform Sandbox Configure on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Platform Sandbox Configure is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Sandbox Lifecycle Management

Manage Salesforce sandbox environments through Connect REST API — list inventory, activate or discard completed refreshes, create and refresh sandboxes, and permanently delete sandboxes.

When This Skill Owns the Task

Use platform-sandbox-configure when the work involves:

  • Listing or retrieving all sandboxes (GET /sandbox/reports)
  • Getting details or status of a specific sandbox by name or by ID (07E prefix)
  • Activating a sandbox after a refresh completes (applying the refresh)
  • Discarding a completed refresh (keeping existing sandbox data unchanged)
  • Permanently deleting a sandbox to free up licenses
  • Verifying activation or deletion completed
  • Creating a new sandbox (Developer, Developer Pro, Partial Copy, or Full)
  • Refreshing an existing sandbox with latest production data

Delegate elsewhere when the user is:

  • Cloning a sandbox → Tooling API (SandboxInfo sObject)
  • Generating a post-copy automation JSON config from an SOP → automation-sandbox-post-copy-config-generate
  • Applying/running a post-copy automation JSON config against a sandbox → automation-sandbox-post-copy-configure

Help (Interactive Menu)

When the user types sandbox -help or sandbox help, respond with ONLY a text message showing the numbered operation list below. Do NOT call any API or tool — just display this menu and wait for the user to reply with a number.

The agent MUST respond with this exact markdown (not in a code block — render it directly as a bullet list):

Sandbox Lifecycle Management

1. Inventory & Details

  • a. List all sandboxes — Names, types, statuses, IDs
  • b. Get details (by name) — Status, license, config
  • c. Get details (by ID) — Provide a 07E ID directly

2. Create & Refresh

  • a. Create a new sandbox — Dev, Dev Pro, Partial, Full
  • b. Refresh a sandbox — Latest production data

3. Activate, Discard & Delete

  • a. Activate a sandbox — Apply a completed refresh
  • b. Discard a refresh — Reject, keep existing data
  • c. Delete a sandbox — Permanent removal

4. Verify & Monitor

  • a. Verify activation status — Check if activate completed
  • b. Verify deletion status — Check if delete completed

5. Post-Copy Automation

  • a. Create Post Copy Automation JSON Configs — Generate a config from an SOP
  • b. Run Post Copy Automation — Apply a config JSON to a sandbox

Reply with a code (e.g. "3a") or describe what you need.

After the user replies, ask for the required input:

SelectionFollow-up question
1aNo input needed — proceed immediately
1b"What's the sandbox name?"
1c"What's the sandbox ID? (starts with 07E)"
2a"What name for the new sandbox, and what license type? (Developer, Developer_Pro, Partial_Copy, Full)"
2b"Which sandbox do you want to refresh? Optionally, provide a new name and/or description."
3a"Which sandbox? Provide a name or 07E ID."
3b"Which sandbox? Provide a name or 07E ID."
3c"Which sandbox? Provide a name or 07E ID."
4a"Which sandbox did you activate? Provide a name or 07E ID."
4b"Which sandbox did you delete? Provide a name or 07E ID."
5a"Delegating to the post-copy config generator — please share the SOP (file, text, or screenshot)."
5b"Delegating to the post-copy config runner — please share the config JSON file and the target sandbox."

Execute the corresponding operation from the Operations section below. Exception: 5a and 5b delegate to a different skill instead of an in-skill operation:

  • 5a → Invoke the automation-sandbox-post-copy-config-generate skill.
  • 5b → Invoke the automation-sandbox-post-copy-configure skill.

This skill does not implement post-copy automation itself — it only routes to the two skills above. Do not attempt to generate or apply a post-copy config directly from this skill.


API Base

CRITICAL: For sandbox operations in this skill, use the Connect REST API. Two discovery paths exist:

  • By name: Call GET /sandbox/reports to list all sandboxes and find the matching one by sandboxName.
  • By ID (07E prefix): Call GET /sandbox/sandboxes/{sandboxId} directly — do NOT call /sandbox/reports.

Both paths return sandbox records with sandboxId (prefix 07E) which is required for all lifecycle mutation operations.

bash
# List all sandboxes (Connect REST API)
sf api request rest "/services/data/v66.0/sandbox/reports" --method GET

# Response format:
# {
#   "count": 3,
#   "sandboxes": [
#     {
#       "sandbox": {
#         "sandboxId": "07E...",        # Required for all operations
#         "sandboxName": "mybox",       # Top-level field — use this for name lookup
#         "license": "Developer",
#         "isPendingActivation": false,
#         "canActivate": true,
#         "canDelete": true,
#         ...
#       }
#     }
#   ]
# }

IMPORTANT: Do NOT use Tooling API (SandboxInfo or SandboxProcess) for sandbox discovery. The mutation endpoints (activate/discard/delete) require the sandboxId (07E prefix) from the Connect REST API response, NOT the SandboxInfo.Id (0GQ prefix) or SandboxProcess.Id (0GR prefix).

NEVER use SOQL / run_soql_query / sf data query for sandbox lifecycle reads — status, inventory, details, license, or pending-activation state (e.g. a "get details / status / license / pending-activation for sandbox X" request). This data lives ONLY in the Connect REST API response (GET /sandbox/reports for name lookup, GET /sandbox/sandboxes/{07E-id} for ID lookup); there is no SObject that returns it correctly. (sf data query --use-tooling-api on SandboxInfo remains valid for the Create and Refresh flows in Operations 7 and 8, which look up the SandboxInfo record to mutate it — that is not a lifecycle read.) If a name lookup returns an empty inventory (count: 0), the sandbox does not exist — report an honest not_found; do NOT retry the lookup via SOQL and do NOT fabricate details.

Report the API result exactly as it comes back — never invent an error or a cause. A count: 0 response is a successful result meaning the sandbox is absent: record it directly as not_found with the endpoint and empty inventory as evidence. Do NOT reinterpret an empty list as an API failure. If the Connect REST API genuinely returns an error, capture that error body verbatim as the outcome — but do NOT speculate about why (e.g. "this must be a scratch org", "sandbox endpoints aren't supported here"). This endpoint does not report the org's edition or type, so any such explanation is a fabrication and must not appear in the output.

Discovery Patterns:

When user provides a sandbox NAME:

  1. Call GET /sandbox/reports to get the full list
  2. Iterate through sandboxes[] array
  3. Check sandbox.sandboxName (top-level field) to find the matching sandbox
  4. Extract sandbox.sandboxId from that record
  5. Use the sandboxId in subsequent mutation operations

When user provides a sandboxId (07E prefix) directly:

  1. Call GET /sandbox/sandboxes/{sandboxId} to verify it exists and check current status
  2. Use the same sandboxId directly in the mutation endpoint — do NOT call /sandbox/reports

Required permission: ManageSandboxes


Operations

1. List Sandbox Inventory

Endpoint: GET /services/data/v66.0/sandbox/reports

Returns a list of all sandboxes with their IDs, names, statuses, and license types.

bash
sf api request rest "/services/data/v66.0/sandbox/reports" --method GET

Response format:

json
{
  "count": 3,
  "sandboxes": [
    {
      "sandbox": {
        "sandboxId": "07E...",
        "sandboxName": "DevBox1",
        "license": "Developer",
        "isPendingActivation": false,
        "canActivate": true,
        "canDelete": true,
        "canDiscard": false
      }
    }
  ]
}

Use when: User asks "show me all sandboxes", "how many sandboxes do I have", "what's the status of my sandboxes"

Key response fields:

  • sandbox.sandboxId (07E prefix) — Required for all mutation operations
  • sandbox.sandboxName — The sandbox name (top-level field)
  • sandbox.license — Developer, Developer Pro, Partial Copy, Full
  • sandbox.isPendingActivation — true if refresh is pending activation
  • sandbox.canActivate / canDelete / canDiscard — Permission flags

2. Get Sandbox Details

Endpoint: GET /services/data/v66.0/sandbox/sandboxes/{sandboxId}

Returns detailed info for a specific sandbox.

Use when: User asks about a specific sandbox's status, configuration, or metadata.

Key response fields:

  • status — Active, Pending Activation, Activating, Completed, etc.
  • isPendingActivation — true if a refresh completed and awaits user decision
  • sandboxType — Developer, DeveloperPro, PartialCopy, Full
  • sourceId — ID of the source org

3. Activate Sandbox (Apply Refresh)

Endpoint: PATCH /services/data/v66.0/sandbox/activate/{sandboxId}

CRITICAL DOMAIN RULE: This operation ONLY applies to sandboxes with a completed refresh in "Pending Activation" state. It applies the refreshed data to the sandbox. It does NOT "bring an inactive sandbox online" or "start" a sandbox.

Pre-conditions:

  • Sandbox must be in Pending Activation status
  • A refresh must have completed successfully
  • User must have ManageSandboxes permission

Before calling PATCH /activate:

  • Confirmed sandbox is in Pending Activation status via GET /sandbox/sandboxes/{id} (isPendingActivation: true)
  • Confirmed a refresh has completed successfully

Use when: User says "activate it", "apply the refresh", "use the latest data"

After activation: The sandbox runs with the newly refreshed production data.


4. Verify Activation

Endpoint: GET /services/data/v66.0/sandbox/sandboxes/{sandboxId}

Poll this endpoint after activation to confirm status changed to Active. This is a verification step, not a standalone user action.

Use when: Agent needs to confirm activation completed (called automatically after activate).


5. Discard Sandbox (Reject Refresh)

Endpoint: DELETE /services/data/v66.0/sandbox/discardsandbox/{sandboxId}

CRITICAL DOMAIN RULE: This operation ONLY applies to sandboxes with a completed refresh in "Pending Activation" state. It rejects the refresh — the existing sandbox continues running with its current data unchanged. It does NOT:

  • Free up licenses
  • Soft-delete or hide the sandbox
  • Reset the sandbox to match production

Pre-conditions:

  • Sandbox must be in Pending Activation status
  • A refresh must have completed

Before calling DELETE /discardsandbox:

  • Confirmed sandbox is in Pending Activation status via GET /sandbox/sandboxes/{id} (isPendingActivation: true)
  • Confirmed this is a discard (reject refresh), NOT a delete (permanent removal)

Use when: User says "discard the refresh", "keep existing data", "don't apply the refresh", "reject the refresh"

WARNING: Discard is not reversible. The user will need to trigger a new refresh if they want fresh production data later.


6. Delete Sandbox (Permanent)

Endpoint: DELETE /services/data/v66.0/sandbox/deletesandbox/{sandboxId}

Permanently removes a sandbox and frees the license.

Pre-conditions:

  • Sandbox must exist
  • User must have ManageSandboxes permission

Before calling DELETE /deletesandbox:

  • Surfaced sandbox details (name, license, status) with user and received explicit delete approval

Use when: User says "delete this sandbox", "remove it permanently", "free up the license"

WARNING: This is irreversible. Always confirm with the user before executing. Surface the sandbox name, license, and status as a safety check.


7. Create a New Sandbox

Creates a new sandbox from scratch. Two approaches are supported — pick based on the user's preference; default to Approach A unless the user asks for a definition file or a repeatable DX blueprint.

Approach A — Tooling API record (direct)

API: Tooling API — SandboxInfo sObject

Required inputs:

  • SandboxName — Name for the new sandbox (alphanumeric, max 10 chars)
  • LicenseType — One of: Developer, Developer_Pro, Partial_Copy, Full

Optional inputs:

  • Description — Description of the sandbox purpose
  • Features — true to upgrade sandbox data storage to 400 MB (WARNING: once enabled, cannot be decreased)
  • ApexClassId — ID of an Apex class that implements SandboxPostCopy interface (runs after creation completes)
  • ActivationUserGroupId — ID of a Group that determines which users can access the sandbox
bash
# Create a Developer sandbox
sf data create record --sobject SandboxInfo --use-tooling-api --values "SandboxName='mybox' LicenseType='Developer'"
Approach B — Sandbox definition file (Salesforce CLI)

The DX-native path: write a JSON definition file (a reusable blueprint), then create the sandbox from it with sf org create sandbox. Prefer this when the user wants a checked-in, repeatable config or name-based Apex/group references (no ID lookups).

json
// config/dev-sandbox-def.json
{
  "sandboxName": "mybox",
  "licenseType": "Developer"
}
bash
sf org create sandbox --definition-file config/dev-sandbox-def.json --alias mybox --target-org prod

Definition file fields:

FieldRequiredNotes
sandboxNameYesAlphanumeric, max 10 chars
licenseTypeYesDeveloper, Developer_Pro, Partial, Full — note: Partial, not Partial_Copy in the definition file
descriptionNoPurpose of the sandbox (≤1000 chars)
apexClassName / apexClassIdNoApex class implementing SandboxPostCopy; the definition file adds the Name variant so no ID lookup is needed
activationUserGroupName / activationUserGroupIdNoPublic group controlling sandbox access; Name variant avoids an ID lookup
featuresNo"['SandboxStorage']" to upgrade data storage (Developer → 400 MB, Dev Pro → 2 GB); not for Partial/Full
templateIdPartial (required), Full (optional)Sandbox template (15-char ID beginning 1ps) selecting which objects to copy
historyDays / copyChatter / copyArchivedActivitiesNoFull sandboxes only

Pre-conditions:

  • Available license of the requested type must exist in the org
  • Sandbox name must be unique and not already in use
  • User must have ManageSandboxes permission

After creation: A SandboxProcess record is created with Status = Processing. The sandbox copy begins immediately.


8. Refresh an Existing Sandbox

Refreshes a sandbox with the latest production data. Two approaches are supported — pick based on the user's preference; default to Approach A unless the user asks for a definition file.

Approach A — Tooling API record (direct)

API: Tooling API — SandboxInfo sObject (PATCH)

Refreshes by updating the existing SandboxInfo record.

Required inputs:

  • Sandbox name — to look up the SandboxInfo record ID (0GQ prefix)

Optional inputs:

  • SandboxName — New name for the refreshed sandbox (if user wants to rename it; alphanumeric, max 10 chars)
  • Description — New or updated description for the sandbox
  • AutoActivate — true to auto-activate when refresh completes (default: false)
  • Features — true to upgrade sandbox data storage to 400 MB (WARNING: once enabled, cannot be decreased)
  • ApexClassId — ID of an Apex class that implements SandboxPostCopy interface (runs after refresh completes)
  • ActivationUserGroupId — ID of a Group that determines which users can access the sandbox

Ask the user: "Which sandbox do you want to refresh? Optionally, provide a new name and/or description if you'd like to change them."

Steps:

bash
# 1. Look up the SandboxInfo record Id by name
sf data query --query "SELECT Id, SandboxName, LicenseType, Description FROM SandboxInfo WHERE SandboxName = '<name>'" --use-tooling-api --json

# 2. Update the record to trigger refresh (PATCH the SandboxInfo record)
# Include SandboxName and Description only if the user provided new values
sf data update record --sobject SandboxInfo --use-tooling-api --record-id <0GQ-id> --values "AutoActivate=true SandboxName='<newName>' Description='<description>'"

Note: Only include SandboxName in --values if the user wants to rename. Only include Description if the user provides one. Always include AutoActivate.

Approach B — Sandbox definition file (Salesforce CLI)

Refresh from the same JSON definition-file blueprint used for create (see Operation 7 for the full field table), using sf org refresh sandbox. Use the existing sandbox's name; the definition file supplies any changed settings (e.g., autoActivate, apexClassName).

json
// config/dev-sandbox-def.json
{
  "sandboxName": "mybox",
  "licenseType": "Developer",
  "autoActivate": true
}
bash
sf org refresh sandbox --name mybox --definition-file config/dev-sandbox-def.json --target-org prod

Pre-conditions:

  • Sandbox must exist and be in a refreshable state
  • Refresh interval must have elapsed (Developer = 1 day, Dev Pro = 1 day, Partial = 5 days, Full = 29 days)
  • User must have ManageSandboxes permission

After refresh: A new SandboxProcess record is created with Status = Processing. If AutoActivate=true, the sandbox activates automatically when done. Otherwise it enters Pending Activation state.


Decision Guide for Agents

When user provides a sandbox NAME (lookup required)

User says...OperationKey check
"Show all my sandboxes"GET /sandbox/reports—
"What's the status of X?"GET /sandbox/reports, filter by sandboxName—
"Activate sandbox sbxtest"1. GET /sandbox/reports to find sandboxId by name2. PATCH /sandbox/activate/{sandboxId}Must be isPendingActivation: true
"Discard the refresh on sbxtest"1. GET /sandbox/reports to find sandboxId by name2. DELETE /sandbox/discardsandbox/{sandboxId}Must be isPendingActivation: true
"Delete sandbox sbxtest"1. GET /sandbox/reports to find sandboxId by name2. DELETE /sandbox/deletesandbox/{sandboxId}Confirm with user first

When user provides a sandboxId (07E prefix) directly

User says...OperationKey check
"Get details for 07E..."GET /sandbox/sandboxes/{sandboxId}—
"Activate sandbox 07E..."1. GET /sandbox/sandboxes/{sandboxId} to verify status2. PATCH /sandbox/activate/{sandboxId}Must be isPendingActivation: true
"Discard refresh on 07E..."1. GET /sandbox/sandboxes/{sandboxId} to verify status2. DELETE /sandbox/discardsandbox/{sandboxId}Must be isPendingActivation: true
"Delete sandbox 07E..."1. GET /sandbox/sandboxes/{sandboxId} to verify existence2. DELETE /sandbox/deletesandbox/{sandboxId}Confirm with user first

Common Mistakes to Avoid

MistakeCorrect understanding
Using activate to "start" any sandboxActivate ONLY applies completed refreshes
Using discard to "hide" or "soft-delete"Discard ONLY rejects a pending refresh
Activating without checking status firstAlways verify isPendingActivation = true
Activating without user confirmationAlways confirm with the user before applying a refresh — this replaces existing sandbox data
Discarding without user confirmationAlways confirm with the user before discarding — this is irreversible and the refresh data is lost
Deleting without user confirmationAlways show sandbox info and ask for explicit confirmation

Frequently asked questions

What does the Platform Sandbox Configure AI skill do?

MUST USE this skill for ANY sandbox request — including simply getting a sandbox's details, status, license type, or pending-activation state by name or ID. TRIGGER when the user: types "sandbox -help"/"sandbox help"; mentions a sandbox ID (07E prefix); asks to list or show sandboxes; asks for the details, status, license type, or config of a sandbox by name or ID; activates or discards a completed refresh; deletes a sandbox; verifies activation or deletion; creates or refreshes a sandbox. DO NOT TRIGGER when: the user wants to clone a sandbox.

Why use Platform Sandbox Configure on TypingMind?

Because you install it once and use it with any model. Platform Sandbox Configure is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Platform Sandbox Configure in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/forcedotcom/sf-skills/tree/main/plugins/builder/dx-org-lifecycle/skills/platform-sandbox-configure. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Platform Sandbox Configure?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Platform Sandbox Configure?

As many as you like. As long as a model supports skills, you can use Platform Sandbox Configure with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Platform Sandbox Configure AI skill free?

Yes. It is published on GitHub by forcedotcom under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇