leadlock-agent-kit

Plugin leadlock-agent-kit from marketplace leadlock in colliercoder/LeadLockAgentKit. Commands read from the repo's manifest — 3 days ago.

Skills (24)

• a2p-registration

  Register your business for A2P 10DLC so your SMS actually gets delivered instead of silently filtered by US carriers. Walks you through all three layers in order - Customer Profile (business identity), Brand Registration (TCR), and Campaign Registration - and fills every field for a one-shot approva

• add-to-learnings

  Capture session learnings into the kit's `learnings/` folder so the entire ecosystem gets smarter over time. Reviews what happened in the current session, surfaces what went wrong, what worked, and what to avoid, then routes each entry into the right category file (bugs, best practices, anti-pattern

• agent-from-recording

  Clone a real salesperson's call patterns into a voice AI agent. Takes a transcript (or audio file you've already transcribed), extracts the call flow, objection responses, personality, and closing patterns, then writes a house-style prompt and creates the agent(s). Triggers on /agent-from-recording,

• agent-to-agent-call-testing

  Functional test harness where two of YOUR Leadlock agents call each other over real Twilio, so call-time features are verified end-to-end against the true production call path (not faked by a simulator or the browser playground). One agent (the "user-sim") dials another (the "agent-under-test"); the

• build-agent

  Build a fully-configured voice agent end-to-end via guided Q&A. Discovers existing phone numbers and GHL calendars on the account, asks for the ~10 inputs that actually matter, picks smart defaults for everything else, generates a house-style system prompt, creates the agent, optionally assigns a ph

• call-audit

  Pull recent calls for an agent, score each one with AI, surface failure patterns, and generate a coaching plan. Like prompt-tuner but read-only — produces a report without touching the prompt. Triggers on /call-audit, audit my calls, audit recent calls, review the calls, review my agent's calls, wee

• clarify-voice

  Turns a rambling voice-to-text note into a short, clean summary of what the user wants, confirms it with them, and only then starts the work (usually by handing off to the matching kit skill). Use when the user dictates to Claude Code or pastes a messy transcript. Triggers on /clarify-voice, clarify

• connect-n8n

  Connect Claude Code to an n8n instance through n8n's built-in MCP server, so Claude can build, edit and run n8n workflows (the backend for Leadlock custom functions) in plain English. Explains what n8n is, n8n Cloud vs self-hosted, enabling Instance-level MCP, connecting with OAuth or a token, and c

• ghl-workflow-api

  Read and write GoHighLevel (GHL) workflows and assets directly through GHL's private backend API using a captured browser session token, no clicking through the builder. List workflows, dump any workflow node-by-node, audit every workflow for broken references (bad pipeline/stage, dead sub-workflow

• install-addon

  Install a paid Leadlock add-on (the Leadlock GHL Connector - GHL CLI with its Claude skill plus the Chrome extension; also in the Launch Bundle) that the user dropped into ./addons/. Installs it, links its Claude skill into the kit, reuses the GHL keys already in .env, and walks through loading the

• install-support-team

  Installs the advanced course's AI support team for "Granite Peak Equipment Rental" into the user's own accounts, end to end - n8n data tables seeded from the kit's CSVs, four n8n workflows (Verify Caller, Rentals, Maintenance, Billing) behind a shared-secret header, three Leadlock agents (Joanne on

• kit-audit

  Point the kit at itself and check it's still tight for the people who clone it — every skill structurally sound, every endpoint real, every skill registered, no hard-rule violations, learnings healthy, no leaked secrets. Static lint plus read-only live probes, auto-fixes safe drift, and writes a sco

• kit-rules

  The Leadlock Agent Kit's operating rules. Load before any Leadlock task (building, tuning, testing or auditing a voice agent, sub-accounts, demos, phone numbers, the Leadlock connector) when the kit is installed as a Claude Code plugin, because a plugin does not load the kit's CLAUDE.md on its own.

• knowledge-base-builder

  Build a knowledge base for an agent from URLs, FAQ documents, CSVs, or markdown — and attach it. Uses the platform's web crawler, FAQ generator, and table parser to assemble searchable Q&A blocks. Triggers on /knowledge-base-builder, build a KB, create a knowledge base, give my agent a knowledge bas

• pick-your-offer

  The pick-your-offer prompt pack. Coaches a new or growing voice-AI agency owner from "I don't know what to sell" to a one-sentence offer, a first setup price and a monthly price in about an hour, using only what Leadlock can deliver today and the four value-equation questions. Asks one question at a

• prompt-tuner

  Tune an agent's system prompt based on real call performance. Pulls recent calls, runs AI analysis, surfaces failure patterns, proposes prompt revisions, and applies them on confirmation. Triggers on /prompt-tuner, tune my agent, improve the prompt, fix my agent, agent is failing, agent is broken, t

• prospect-demo

  Build voice agent demo(s) for a prospect from a URL or filled-out demo intake. Scrapes the site, writes the system prompt, creates 1-to-N agents across voice models (Gemini, OpenAI, Grok, ElevenLabs) for A/B testing, creates orb demo configs, and returns live shareable URLs ready to embed in any GoH

• retell-to-leadlock

  Port a Retell AI agent (or a Retell orchestrator + its transfer/sub-agents) into an equivalent Leadlock agent. Pulls the Retell LLM definitions, maps the prompt + every tool, consolidates a multi-agent transfer tree into ONE Leadlock prompt (Leadlock has no agent-to-agent transfer yet), recreates cu

• sales-presentation

  Turn a single fill-in template into a branded, client-facing sales deck that pitches an AI voice-agent system to any business in any industry. Produces one standalone dark-themed HTML page (single-scroll, not a slideshow) built from a reusable component library - hero with badges, numbered sections,

• self-selling-setup

  Walk an agency step-by-step through standing up the Leadlock self-selling GHL snapshot for their own business - from imported snapshot to a live, paying, hands-off system. Claude runs an intake interview, auto-fills every GHL custom value via the public API, hands over an ordered checklist for the G

• setup-check

  Verify the kit is configured correctly. Probes the .env file, tests the API key against the Leadlock API, identifies the user's tenant tier (agency or sub-account), and reports which skills they can run. Use proactively when ANY other skill fails with a 401, when the user reports "it's not working",

• subaccount-onboarding

  Onboard a new agency client end-to-end — create the sub-account, set pricing economics, wire integrations, copy a template agent, and send the welcome invite. Compresses the 30-minute UI workflow into one command. Requires an agency-tier API key. Triggers on /subaccount-onboarding, onboard a new cli

• talking-website

  Turn any website into a talking website — a voice agent the visitor clicks and speaks to, that controls the page (navigates, scrolls, jumps to a section, spotlights an element), books appointments on the account's calendar, captures the lead into the CRM, and logs every conversation in Leadlock. Rou

• welcome

  Show the user what skills are available in this kit and example phrases that invoke each one. Use proactively when a user is new to the kit, asks "what can this do", "help", "what's available", "I'm new", "where do I start", "list skills", "show me what you can do", "/help", "how do I use this", "/w

MCP servers (1)

• leadlock

  connects to: https://mcp.leadlock.ai/mcp

Hooks (1)

• PreToolUse

  python3 "${CLAUDE_PLUGIN_ROOT}/.claude/hooks/guard_secrets.py"

Similar plugins