clara
INSURANCE.md standard steward

Insurance for the agent era.

Clara helps businesses build and own their insurance and risk context by turning scattered facts, contracts, controls, policies, claims history, and evidence into INSURANCE.md: a portable readiness packet their own agents can maintain.

Explore INSURANCE.md
INSURANCE.mdSTRUCTURED
companyAcme Agents, Inc.
purposePrepare risk context before renewal, diligence, or advisor review.
evidenceSOC 2, customer MSA, current policies, security controls.
gapsRevenue split, AI use, claims/circumstances confirmation.
ownerCompany-controlled packet; human review before third-party use.

Insurance context should belong to the business.

Insurance is still a risk-transfer market, but the preparation layer should not live only in renewal PDFs, inboxes, third-party portals, or someone else's memory. Clara makes the context structured, portable, and useful to agents before any external review begins.

Protocol

INSURANCE.md gives every company a portable, evidence-linked packet for risk facts, contracts, controls, policies, claims history, and open questions.

Skills

Agents can generate and maintain the packet inside Codex, Claude Code, Cursor, OpenClaw, Antigravity, or another company workbench.

Bench

Clara Bench measures whether agents are preparing risk context with evidence, uncertainty labels, boundary discipline, and reviewable outputs.

Workspace

The hosted workspace turns the file into company-owned memory: saved packets, validation history, evidence checks, renewal diffs, and exports.

Advisory

Readiness sprints help teams deploy the framework, stand up internal risk-agent workflows, and organize evidence without buying transaction services.

Intelligence

Aggregate readiness patterns, benchmark results, and control libraries can power better recommendations while keeping private company data protected.

Agent-readable by default

Designed for workbenches, not sales funnels.

If a founder points Codex, Claude Code, Cursor, OpenClaw, or another harness at Clara, the agent should understand what to generate, what evidence to inspect, what not to claim, and what needs human authorization.

INSURANCE.md structure
---
schema_version: "clara-insurance-md-v0"
company: "Acme Agents, Inc."
generated_at: "2026-06-01"
prepared_by: "company_agent"
readiness_state: "STRUCTURED"
attestation_status: "NOT_REVIEWED"
overall_confidence: "MEDIUM_CONFIDENCE"
---

# INSURANCE.md

## Business Summary
Acme Agents builds operational AI agents for finance and security teams.

## Insurance Requirements
- Enterprise customer contract requires Tech E&O, cyber, CGL, and D&O.
- Cyber minimum: $2M aggregate.
- Contract asks for notice of cancellation within 30 days.

## Evidence Index
- SOC 2 Type II report: evidence/security/soc2-2026.pdf
- Customer MSA: contracts/customer-msa-acme-bank.pdf
- Current cyber policy: insurance/cyber-policy-2025.pdf

## Gaps
- Confirm revenue split by geography.
- Confirm whether any model outputs are used for regulated decisions.
- Confirm known claims, incidents, or circumstances.

## Human Review
Status: NOT_REVIEWED
Reviewer: pending

The Clara readiness loop.

The first wedge is not an automatic insurance transaction. It is a structured operating loop that helps the company and its agents understand risk before decisions get outsourced.

From scattered inputs to readiness memory

  • Collect authorized documents and risk facts
  • Generate INSURANCE.md locally or in Workspace
  • Validate evidence, gaps, and unsupported claims
  • Benchmark readiness quality against Clara rubrics
  • Export an advisor-ready packet the company controls

What gets standardized

  • Company profile
  • Operations and revenue
  • Products and services
  • AI, model, and data usage
  • Contracts and insurance requirements
  • Current policies
  • Controls and evidence
  • Claims and circumstances
Company-owned data

A risk manager in a box.

Workspace is the bridge for teams that are not fully agent-native yet. Upload documents, answer missing questions, generate the packet, track changes, and export summaries your team can use with existing advisors.

Open Clara workspace

Workspace modules

  • Document intake and evidence index
  • INSURANCE.md generator and validator
  • Gap list, renewal diff, and review checklist
  • Advisor-ready executive exports
  • Premium recommendations from Clara Intelligence
Clara Intelligence

Benchmarks become the proof layer.

Clara Bench can test whether agents extract facts, preserve uncertainty, cite evidence, avoid unsupported advice, and produce outputs a human can actually review. Over time, aggregate patterns can help companies understand what organizations like them should have in place.

Readiness reports

A trust-building layer that helps Clara publish useful research while improving the protocol and workspace.

Control libraries

A trust-building layer that helps Clara publish useful research while improving the protocol and workspace.

Peer patterns

A trust-building layer that helps Clara publish useful research while improving the protocol and workspace.

Agent evals

A trust-building layer that helps Clara publish useful research while improving the protocol and workspace.

Future platform

Monetize the workflow, not the public standard.

Clara can keep the standard open while building paid workspace features around saved state, validation, exports, collaboration, benchmark reports, premium recommendations, and readiness sprints.

Near-term business model

  • Free public protocol and open skills
  • Freemium workspace for company-owned readiness memory
  • Paid validation, exports, benchmark reports, and team history
  • Advisory readiness sprints for higher-touch implementation
  • Intelligence modules built from aggregate, privacy-safe patterns
Explore Clara Bench