Agentman
Bring Joy Back to Your Practice
🛡️ AgentReady threat assessment
MAESTRO 7-layer threat model + OWASP AIVSS risk score for Agentman, derived from its capabilities.
These scores are auto-generated from public information (the agent's own listing, docs, and repository) using the canonical OWASP AIVSS formula and the MAESTRO framework — an estimate for guidance, not a penetration test, audit, or certification. See the scoring methodology — every score is re-derived by the same automated method as an agent's public evidence changes.
Overview
Agentman turns your team's expertise into agent skills — reusable, versioned procedures that run identically across Claude, ChatGPT, Cursor, Claude Code, and Agentman's own agents through a single MCP connection. Built once, shared across the team, managed in one place, with the permissions, versioning, and audit trails that legal, private equity, and healthcare work demand. Three ways to ship: build skills yourself (free to start), build a custom agent harness with Agentman, or deploy a fully built product like Medman — a complete medical back office live in clinics today. SOC 2 Ready, ISO 27001, and HIPAA. Built by operators out of Salesforce Frontier AI, Rappi's GenAI engineering org, and Walmart Connect, alongside a practicing clinician whose own practice runs on Agentman.
Key features and capabilities
- Skills, not prompts — Procedures, decision logic, thresholds, and edge cases packaged as versioned, permissioned SKILL.md files. Update once; every agent and teammate gets it.
- One connection, every tool — The same skill runs in Claude, ChatGPT, Cursor, Claude Code, and Agentman agents over MCP. No per-platform rebuilds.
- Skills → Agents → Harness — A skill is the unit of work, an agent runs a single loop, a harness orchestrates hundreds for hours — with cross-session state, so a failed step resumes instead of restarting.
- Governance built in — Role-based permissions, draft → review → publish workflows, exportable audit trails, one-click version rollback, confidentiality walls by client or matter, and human-in-the-loop on consequential decisions.
- Curated skills library — 87+ practitioner-authored skills across 22 categories — healthcare RCM, legal, PE, sales, content ops — cloneable and customizable on day one.
- Three ways to ship — Built for you (Medman, live in clinics), build with us (Agentic Harnessing), or build yourself (Agent Skills + no-code Agent Builder, free to start).
Use cases
- Healthcare (Medman) — A complete medical back office on Agentman skills: eligibility verification across 2,700+ payers, inbox triage, and claims loops with a full audit trail, no EHR integration required. Live: 90% eligibility automation in 12 weeks, 65% fewer denials, 50→7 monthly verification hours, $0.50/check vs. the $6.72 CAQH benchmark. Two agents live, six on the 2026 roadmap.
- Legal — Contract review and redlining against your standards, clause extraction, NDA triage, and compliance checks, scoped per client and matter.
- Private equity — Full-data-room diligence, CIM analysis, IC memo generation, and deal screening — orchestrated as a harness that resumes on failure.
- Codify expert workflows — Turn the playbooks in your best people's heads — review frameworks, qualification criteria, payer rules — into skills any tool or teammate runs consistently.
- Sales & marketing — Discovery frameworks, qualification logic, campaign planning, and competitive tracking — your methodology applied consistently across the team.
- Custom high-stakes automation (Agentic Harnessing) — When the job exceeds a single loop, Agentman builds a production harness coordinating hundreds of skill executions with audit trails and governance — the architecture proven in Medman, applied to your firm.