Skills Index
Every skill the APIANT plugin exposes to Claude Code, grouped by workflow. Claude picks the right one from your prompt — you rarely type the slash command.
You don't type these commands. Claude Code reads your prompt and auto-invokes the right skill. This page is a catalog so you can see what's in the toolbox; the path you actually use is plain English.
The complete inventory of plugin skills Claude Code can run against APIANT. Every skill is a procedure a senior APIANT engineer follows, encoded as a slash command. Click any card for the full what/when/how.
Verify MCP connections, environment, and plugin version. Run on first install or when something breaks.
The integration plumbing: OAuth, API Key, or no-cred connectors, triggers, and actions (ADD/GET/LIST/UPDATE/DELETE/FIND).
Multi-automation deployments: two-way sync, data pipelines, coordinated workflows with shared state.
Form schema for lead capture, onboarding, or config, embedded straight into an APIANT automation.
Add a step, fix the mapping, swap Mailchimp for Klaviyo, disable a branch. Surgical, non-destructive edits.
Fix JSP, update API calls, change endpoints, rename settings, edit dropdowns, with full diff awareness.
Execute, inspect results, diagnose failures, apply fixes, retest, and run branch coverage automatically.
Validate data flow between automations, shared state, loop prevention, and cross-automation coordination.
Publish template automations and roll them out to customer accounts in one coordinated step.
Poll an endpoint on a schedule and fire when a new record appears. The workhorse trigger.
Polls and fires when a record is created or modified. Tracks last-seen cursor automatically.
You paste a URL into the source app. Incoming posts fire the automation.
APIANT creates and tears down the webhook automatically via the target app's API. Zero manual setup.
Credential-based webhook URL with event filtering. Multiple event types route on a single endpoint.
APIANT-to-APIANT internal protocol. Enables chained automations across accounts and environments.
Create a new record in the target app. Handles validation, required fields, and response parsing.
Remove a record by ID or criteria. Safe-mode defaults with rollback metadata preserved.
Search by criteria, return matches. Used for dedup, upsert, and enrichment patterns.
Retrieve one record by ID. The simplest read path. Often chained after FIND or NEW.
RPC-style POST that returns rich data: AI inference, generation, computation, transformation.
Page through results with cursor or offset handling. Rate-limit aware by default.
Patch existing records. Partial updates supported. Only touches the fields you specify.
OAuth V2, API Key, or no-credential auth. Handles token refresh, scopes, and revocation.
Drive a dev browser to register an OAuth app on a vendor portal, capture client_id and secret into the keyvault.
Migrate legacy human-built assemblies to Claude-template pattern so Claude can safely edit them.
Builds the matching trigger and action pair for a two-way sync. Handles loop prevention, last-write-wins, and conflict resolution out of the box.
From simple chatbots to multi-goal conversation flows with memory and tool-calling automations.
Aggregate items into named buckets across executions, then drain on a schedule. Daily digests, batched summaries.
Parent/child flows with query-string parameter passing and webhook payload forwarding.
Pause an automation until a person approves or denies via moderation queue link. Decision gates for AI flows.
Launch N child automations in parallel, wait for all to finish, proceed with aggregated results.
Expose an automation as a callable tool to Claude Desktop, Cursor, ChatGPT MCP, and other MCP clients.
Pause an automation until a specific datetime. Useful for drip sequences and scheduled follow-ups.
Synchronous HTTP endpoint. Caller waits, automation runs, response comes back.
Cookbook of small one-action utilities: RSS feeds, PDF generation, daily CSV datafiles.
Watches prod automations for errors. Designed for /loop recurring runs. Alerts with full trace context.
Control which errors retry, which 401s bypass auto-turnoff, alert text rewriting, dismissal lists.