AI Sales Agent

September 17, 2026 · View on GitHub

An AI agent for sales — lead generation, ad-based prospecting, LinkedIn outreach, company enrichment, and email verification — backed by real B2B data APIs.

Part of Agency Agents OS, an open ecosystem of specialized AI agents for real business work.

Agency Agents OS — the central catalog this repo is part of.
ai-research-agent — deeper web research on a prospect/company beyond enrichment lookups.
ai-marketing-agent — nurtures leads this repo generates.
MuAPI MCP docs — connect this repo's SKILL.md files via MCP.
MuAPI Agent Skills — background on the SKILL.md pattern this repo uses.
MuAPI access keys — create the API key this agent needs.

What this covers

This repo is the umbrella for anything an agency or in-house sales team would call "the AI sales agent": turning an ideal-customer-profile description into a verified, enriched prospect list, and drafting the outreach to reach them — without hand-operating a separate tool for each step.

Sub-agents

Agent	Does	Status
Lead Generation	Build a targeted prospect list from an ICP description	Tested
Ad-Based Prospecting	Find companies actively running ads as a warm, budget-confirmed lead signal	Blueprint
LinkedIn Outreach	Draft and sequence personalized connection/outreach messages (never auto-sends)	Tested
Company Enrichment	Enrich a company name/domain with firmographic data	Tested
Email Verification	Validate a list of email addresses for deliverability before a campaign	Tested

Required Muapi APIs

company.enrich — firmographic enrichment (size, funding, tech stack, industry) for a company name or domain.
people.search — prospect/contact search against an ICP description.
email.verify — email deliverability validation for a list of addresses.
outreach.draft_message — personalized outreach/connection message drafting.
ads.ad_library_search — public ad-transparency library lookup, used by Ad-Based Prospecting to find companies with active ad spend (shared with ai-ads-agent/ai-competitor-intelligence-agent).
company.technographics — detect a company's website tech stack, or find companies using a given technology (reverse lookup). (coded 2026-09-17, not yet live)
company.funding — a company's funding rounds and financing events, or the latest rounds across companies. (coded 2026-09-17, not yet live)
company.buying_signals — buying/intent signals detected for a company. (coded 2026-09-17, not yet live)
company.job_postings / company.headcount_growth — hiring activity and headcount growth as buying-signal/timing filters. (coded 2026-09-17, not yet live)
people.rank_decision_makers — rank a company's decision-makers to surface the best-fit buyer contact. (coded 2026-09-17, not yet live)
linkedin.company_profile / linkedin.people_search / linkedin.employees — read-only LinkedIn company page, member profile, and employee-roster lookups. (coded 2026-09-17, not yet live)

See each sub-agent's SKILL.md for the specific capabilities it uses.

Setup

Create a Muapi account and API key at muapi.ai.
Review the Muapi API quickstart and OpenAPI schema for the relevant endpoints as they become available.
Load the SKILL.md for the sub-agent you need into your agent runtime (hosted agent, MCP client, or custom LLM app), or follow it manually.

Using with an AI agent

Every sub-agent's SKILL.md is model- and runtime-agnostic — it's plain Markdown, so it works with any LLM agent, not just Claude. Two integration paths:

As an MCP connection (the agent gets live Muapi tools):

Muapi runs an MCP server at https://api.muapi.ai/mcp that any MCP-compatible client can connect to — Cursor, Windsurf, Claude, or your own custom agent.

Cursor / Windsurf / other clients with a header field: connect to https://api.muapi.ai/mcp with an Authorization: Bearer YOUR_MUAPI_KEY header.
claude.ai / Claude Cowork / other connector UIs with no header field: use the URL-embedded key form instead, https://api.muapi.ai/mcp/YOUR_MUAPI_KEY, via Settings → Connectors → Add custom connector.
Claude Code / Claude Desktop: claude mcp add muapi -e MUAPI_API_KEY=YOUR_MUAPI_KEY -- muapi mcp serve (uses the muapi CLI's stdio transport — Claude Code's HTTP MCP client doesn't reliably inject tools).

Full setup details for every client: muapi.ai/docs/mcp.

As agent instructions (any LLM follows the workflow directly):

Drop a sub-agent's SKILL.md into a Claude Code project's .claude/skills/ directory, paste it into a custom-GPT/Project's system instructions, hand it to an autonomous agent framework as a tool spec, or attach it directly in a chat conversation — then ask the agent to follow it.

Read-only vs. write actions

Lead Generation, Ad-Based Prospecting, Company Enrichment, and Email Verification are read-only — they look up, source, and validate data, nothing is sent anywhere. LinkedIn Outreach is requires-approval-to-publish: the agent only drafts connection/outreach messages and sequences; it never sends a message on its own. A human must review and approve every send.

Status and limitations

Four sub-agents are Tested (2026-09-09): company.enrich, people.search, and email.verify are live on Muapi's production API and have each been run end-to-end with real inputs (e.g. company.enrich against stripe.com returned a full firmographic profile; people.search and email.verify returned real matches). LinkedIn Outreach's drafting step is agent-side (no dedicated API beyond the company.enrich lookup it uses for personalization) and is exercised the same way. Ad-Based Prospecting is Blueprint (added 2026-09-10): ads.ad_library_search is coded but only its Meta leg is confirmed working — Google/TikTok/LinkedIn aren't yet functional (same aggregator vendor-key blocker tracked in ai-competitor-intelligence-agent/ai-ads-agent); company.enrich used within this sub-agent is already live and tested.

2026-09-17 addition, not yet reflected in the Tested status above: Lead Generation, Company Enrichment, and LinkedIn Outreach now also reference company.technographics, company.funding, company.buying_signals, company.job_postings, company.headcount_growth, people.rank_decision_makers, and linkedin.company_profile/linkedin.people_search/linkedin.employees in their workflows. These are code-complete server-side but not yet live on Muapi's production API (no DB sync/seed_skills run yet, and several endpoint ids are still unverified best-effort guesses pending a live token call) — each affected sub-agent's own SKILL.md states this plainly. The pre-existing Tested capabilities on these same sub-agents (company.enrich, people.search, outreach.draft_message) are unaffected and remain live.

Contributing

See Agency Agents OS CONTRIBUTING.md.

License

MIT