From Voximplant AI Agent Skills
Automate Voximplant account setup and operations through the Management API, official API clients, or VoxEngine CI. Use when creating or updating applications, users, scenarios, routing rules, phone numbers, call lists, service accounts, sessions, logs, recordings, or deploying VoxEngine code.
How this skill is triggered — by the user, by Claude, or both
Slash command
/voximplant-ai-agent-skills:voximplant-management-apiThe summary Claude sees in its skill listing — used to decide when to auto-load this skill
Use this skill for platform automation, not for writing scenario logic. For VoxEngine scenario code, use `voximplant-voxengine-dev`.
Use this skill for platform automation, not for writing scenario logic. For VoxEngine scenario code, use voximplant-voxengine-dev.
Many real tasks require both skills: this skill retrieves platform state, call history, logs, recordings, and manages resources/secrets; voximplant-voxengine-dev analyzes and fixes scenario code.
https://docs.voximplant.ai/llms.txthttps://docs.voximplant.ai/platform/management-api/llms.txthttps://docs.voximplant.ai/platform/management-api/authorization.mdA2PSendSms and A2PGetSmsHistory may require SMS-specific permissions beyond a general Developer role.--dry-run before uploading.SetScenarioInfo.BindScenario, then verify the binding.SetRuleInfo can rebind a rule to a scenario. It may accept scenario_id and return success without changing the binding.At the beginning of a user request, quickly classify the user's experience level:
Use wording like:
If you are new to Voximplant, I can first explain the moving parts and propose safe defaults. Tell me what call flow you want, and I will say which credentials and approvals are needed.
When the user asks you to read this file or pastes this skill into chat, immediately follow it for Voximplant platform automation. Do not only summarize it.
Tell the user this skill can create and manage Voximplant platform resources for them after they provide service account JSON credentials and approve changes. Do not list applications, rules, users, scenarios, or secrets as things the user must manually create unless the user explicitly wants to do it manually.
If the task also needs VoxEngine code, read or request voximplant-voxengine-dev. If the user has not installed these skills, offer to copy both skill folders into their agent environment after permission:
~/.cursor/skills/When API access is needed, first ask which local project folder is in scope. Do not invent paths.
Then tell the user:
voximplant-service-account.json. If the file keeps its default _private.json suffix, the agent should be able to find it by name. Do not commit it.Prefer the simple root-file handoff for beginners. The agent must proactively check or update .gitignore so the JSON is not committed; create .gitignore if needed. Do not ask the user to paste their main account password, browser cookies, or control panel session tokens.
Do not read the service account JSON contents into chat by default. Pass only the file path to local commands or scripts. Reading the JSON into the agent conversation may send the private key through the LLM provider and store it in conversation logs.
If the user already works with environment variables or CI secrets, adapt to that setup. Do not make a beginner manually edit paths in .env; the root JSON filename is the simplest default.
When discovering credentials, only inspect filenames at first:
*_private.json.If the user is new, offer a guided path:
I can choose safe defaults and do the platform setup for you. I only need your Voximplant service account JSON file name, approval before changing resources, and external provider keys only if your use case needs them.
If the user is deploying scenarios with VoxEngine CI, ask for or configure only when CI is the selected path:
VOX_CI_CREDENTIALS: path to the service account JSONVOX_CI_ROOT_PATH: optional local root for downloaded VoxEngine filesBefore performing Management API work:
Switch to or request voximplant-voxengine-dev when the task requires scenario authoring, code review, runtime event ordering, or code-level debugging. If the user does not have that skill, ask them to provide it or ask permission to fetch the public skill package after it is published:
https://github.com/voximplant/voximplant-ai-agent-skillsKeep the boundary clear:
voximplant-voxengine-dev writes code, validates VoxEngine APIs, uses already available secrets inside scenario code, and interprets logs to fix scenario behavior.After deployment, offer a test call. If the user has no phone number, offer a no-number test using https://phone.voximplant.com/:
.env only after the user approves, or print the variable names and ask the user to fill values.Enter destination: for a call inside the same application, any value is enough, for example 1.When the user reports a failed call:
When the task involves buying or renting phone numbers, warn early that some countries and number types require identity, business, or address verification before the number can be purchased or activated.
Do not imply that the Management API can bypass regulatory verification. Instead:
https://phone.voximplant.com/ while verification is pending.Prefer official tooling over hand-rolled wrappers, with a Management API-first default:
@voximplant/apiclient-nodejsvoximplant-apiclientgithub.com/voximplant/apiclient-govoximplant/apiclient-phphttps://docs.voximplant.ai/platform/management-api/llms.txthttps://docs.voximplant.ai/platform/management-api/overview.mdhttps://docs.voximplant.ai/platform/management-api/basics.mdhttps://docs.voximplant.ai/platform/management-api/authorization.mdhttps://docs.voximplant.ai/platform/management-api/callbacks.mdhttps://docs.voximplant.ai/platform/management-api/child-accounts.mdhttps://docs.voximplant.ai/platform/management-api/secure-objects.mdhttps://docs.voximplant.ai/platform/voxengine/ci.mdFor fuller guidance, read reference.md. For task patterns, read examples.md.
Guides creation, editing, and verification of skills for AI coding agents using test-driven development with subagent scenarios. Use when authoring or debugging skills.
npx claudepluginhub voximplant/voximplant-ai-agent-skills --plugin voximplant-ai-agent-skills