Skills for every AI tool

Teach Claude and ChatGPT how your team works.

A skill captures how your best people actually do the work — their procedures, judgment, and standards — then makes Claude, ChatGPT, and every agent you run do it the same way, every time. Built once, shared across your team, managed in one place.

Works in Claude · ChatGPT · Cursor · Claude Code · Agentman agents — through one connection.
contract-negotiation-playbookLegal · v2.4.0
What a skill knows that the internet doesn't
ProceduresYour review framework by contract type — MSA, SOW, NDA
Decision logicPreferred → acceptable → walk-away, per clause
ThresholdsLiability below 2× fees = partner review
Edge casesThe exceptions only your senior people know
Versioned · permissioned · runs the same in every tool

Trusted by Teams at

Stanford UniversityRappiCiscoNested BeanHandshake

What is Agentman?

Agentman is a platform for turning your team's expertise into agent skills — then putting them to work, from a single free skill to a system that runs an entire operation.

Skills are the raw material. A harness is the machine. Medman is the proof.

Skills first

Everything Agentman does runs on your skills.

A skill is your expertise, codified. An agent executes one. A harness orchestrates hundreds — all drawing from the same library your team owns and improves.

01 · SKILLS

Your expertise, codified

The playbooks, evaluators, and thresholds your best people use — versioned and shared.

This is AgentSkills. Start free, today.
02 · AGENTS

Single loops that execute

Review one contract. Verify one claim. Short-lived, focused, consistent.

Run them in any AI tool you already use.
03 · HARNESS

Orchestration for the whole job

Hundreds of coordinated loops running for hours — with audit trails and cross-session state, so a failed step resumes instead of restarting.

Built with you: Agentic Harnessing.
Centralize your intelligence

One source of truth your whole team — and every AI — works from.

Your team's best work already lives in playbooks, habits, and a few people's heads. A skill brings it into one place every AI tool and teammate can draw from — and improve over time.

Shared, not siloed

Expertise lives in the team library, not one person's chat history.

Versioned

Improve a procedure once; every agent and teammate gets the update.

Permissioned

Control who can view, edit, and publish — scoped by team, client, or matter.

See the difference context makes

Same request. One without your context, one with it.

Generic AI

Contract Review

Review contracts carefully for liability, payment terms, and termination clauses. Ensure all parties understand their obligations and consider consulting an attorney before signing.

Generic. Textbook. Could be anyone's.
AI + your skill

Contract Review

AI + contract-negotiation-playbook
IndemnificationMutual clause confirmed — no unilateral exposure
Auto-renewal60-day window vs. your 90-day standard — redline
Liability cap1× fees — below firm floor (2×); escalate to partner
Specific. Actionable. Actually yours.

A skill reviews one contract against your standards. A harness reviews the entire data room — and picks up where it left off when something fails.

Proof, in production

Medman is what this looks like fully built.

A complete medical back office running on Agentman skills — a production harness coordinating eligibility, inbox triage, and claims loops across 2,700+ payers, with the audit trail to show what every agent did. What Medman is for medicine, Agentic Harnessing builds for your firm.

“We cut 50 hours of monthly verification calls down to 7. Staff aren't drowning in eligibility calls anymore — they're actually talking to patients.”

— Sachin Gangupantula, VP Agentic Healthcare & Practice Operator
90%
eligibility automation in 12 weeks
65%
fewer eligibility denials
50→7
hrs/mo of verification calls
$0.50
per eligibility check vs. $6.72 CAQH benchmark
2
agents live, 6 on the 2026 roadmap
See Medman run a back office
Governance & permissions

Built for firms that can't get this wrong.

Permission management and auditing designed for legal, private equity, and healthcare — where every decision needs a paper trail.

Role-based permissions

Scoped by team, client, matter, or deal — who can view, edit, and publish.

Approval workflows

Draft → review → published. Nothing reaches your AI tools without sign-off.

Full audit trail

Who changed what, when, and which agent used which version on which task. Exportable.

Version history & rollback

Every revision kept. Roll back a procedure in one click.

Confidentiality boundaries

Wall off skills by client or engagement so context never crosses lines.

Human in the loop

Consequential decisions wait for a person. Always reviewable.

SOC 2 ReadyISO 27001HIPAA

Built by operators, not observers

This is a team that has already built, scaled, and operated exactly what Agentman requires.

Prasad Thammineni
Prasad Thammineni
Founder & CEO
Ex-VP Frontier AI, Salesforce. Serial entrepreneur.
Moshe Ojeda
Moshe Ojeda
Co-Founder & Head of Agentic Engineering
Ex-Head of Gen AI Engineering, Rappi. Production AI at scale.
Sachin Gangupantula
Sachin Gangupantula
VP, Agentic Healthcare
9 years running practice operations. His clinic runs on Agentman.
Debby Wang
Debby Wang
VP, Agentic GTM
Ex-Sr Director of BizOps, Walmart Connect. Led #2 U.S. retail media network to +26% revenue growth.