<!-- Generated from archive/home-2026-11-16.html by scripts/site/generate.mjs. Edit the page, not this file. -->

# RiskMandate — Archive: the home page until 16 November 2026

An Agent Behaviour Policy writes down what one AI agent can really reach, what you authorised it to do and the gap between the two. Behaviour policies for the agents in Gmail, HR, payroll, CRM and meeting apps: pick a system, set the rules, and know the moment an agent steps out of line.

Source: https://riskmandate.ai/archive/home-2026-11-16.html

---

# Give agents access.Not free rein.

Agent Behaviour Policies for the AI agents in Gmail, HR, payroll, CRM and meeting apps: what each one can reach, what you authorised it to do, and the gap between the two. Enforce them with the controls you already run, or with [SHarness](/secure-harness.html), the open-source secure harness we maintain. We are never in the request path.

On Claude or ChatGPT, in your browser, nothing to install. [Six examples, simplest first](https://customers.riskmandate.ai/).

## Know their reach. Define their mandate. Control the gap.

An agent, everything it can reach, the mandate the business gave it, and the gap between the two. Then the same record for every agent, read by operations, leadership, security and the insurer. No narration needed.

Film by Silvio Switha, November 2026, also on the [Web Summit site](https://web-summit.sgit.ai/videos/reach-mandate-gap/index.html). The film says delta where this site says gap: the same thing.

Blocked by your gateway: Payroll copilot tried to change bank details for 1 employee. Reply 1 to review.

The gate in the middle is **your** gateway, proxy or IAM, or SHarness on the machine the agent runs on, configured from the rules in the ABP. RiskMandate writes the rules and is never in the request path; the approval card is a flow your team stands up with our help.

## Put a behaviour policy on your agent.

Ask your assistant to do something. Paste one line that says it must not. Ask again, and watch it refuse. That line is the first row of an Agent Behaviour Policy.

In each example the assistant keeps the line because it chooses to: an expectation, the weakest of the four barriers. The examples show why the stronger ones matter.

- _1_**Ask**In Claude or ChatGPT, ask for something you would not want an agent to do on its own.
- _2_**Add the line**Paste one sentence that says it must not.
- _3_**New chat**Ask again, and see what changed, and what did not.

One policy for each system your team already uses

- Gmail
- Google Workspace
- HR
- Payroll
- CRM
- Marketing
- Note takers
- Meeting apps

## Pick a system. Set the rules.

Every system starts from a template. Decide what agents can do, what waits for approval and what never happens. We write it down; your controls enforce it. The rules you can paste into an agent today, one worry at a time, are [here](/rules/).

## Bring your policy. We show what can enforce it.

Upload the document legal already wrote. We turn every line into a rule, supply the business logic behind it, name the control in your stack that would enforce each one, and flag what nothing can enforce as written.

1Staff must not use AI tools to send confidential data outside the company.

2Customer data exports need a manager's approval.

3AI note takers may not record HR or legal meetings.

4Use AI responsibly and ethically at all times.

## A prompt asks. A boundary enforces.

Instructions inside an agent can be ignored or talked around. A control outside it cannot, and the ABP says which of the two each rule has.

"Never export customer lists without approval."

## Enforcement, when you want it.

Every rule in an ABP is kept by something. Told, it is a hope; checked, it is a hook on the agent's own machine; held, it is a boundary the agent cannot reach. **SHarness** is how a row gets held: an open-source secure harness, which we maintain, that you run in your own environment. We are still never in the request path.

### A line in the prompt

The model keeps it because it chooses to. An expectation, and the one the five-minute examples show.

### A hook

Deterministic, and on the agent's own machine, where the same account can switch it off. A setting.

### A boundary

The ABP becomes a Process Behaviour Policy, named by a hash you supply, and the kernel holds it on every process the agent starts. A tricked agent cannot widen it.

_**Watch the kernel refuse.** `sharness-ai hello` writes a PBP, hashes it, lets the kernel hold it, then runs six probes against it: six of six held, 25 October 2026, for a child process. The launcher that would hold the agent itself is still the design. Three more animations, from the gap to the radius, are on [the SHarness page](/secure-harness.html#watch). _E11, from the-ultimate-insider.presentation.sgit.ai, CC BY 4.0, copied here ([credits](/assets/anim/README.md))._Replay[Full screen ↗](/assets/anim/studio/lab/e11/v1/index.html)_

- The kernel locks on a process: files, execute, ports, no network, caps, hash checked.
- `hello`: six probes, six held, in five environments, and in a Linux kernel in a browser tab.
- The file and network doors, as early versions; a shell whose only way out is the network door.
- The launcher above the agent: nothing yet binds the agent itself.
- The key, shell and model doors; seccomp; hostnames in the policy.
- Linux only; no protection against root. Translations to Docker, bubblewrap and Kubernetes are written, not yet run against them.

## Know the moment an agent steps out of line.

An approval flow your team stands up with our help, from the rules in the ABP: the right person gets a text and an email, and approves or denies with a one-letter reply. It runs in your stack, not ours.

Held: Sales agent (for Dana R.) wants to export 2,400 contacts. Your CRM policy holds exports over 500. Reply A to approve once or D to deny.

D

Denied. Dana and the agent have been told why. Decision 48213 is in your log.

## Four levels. One document.

What is for sale is an Agent Behaviour Policy for the agent you actually run. Each level is the level below plus exactly one thing, and the one thing is what the price is for. Pick a level to see what arrives.

Prices are in pounds, per agent, per deployment, paid once at the store; no subscription. These four are what we can deliver today. Level 3 is where most people start, and it is the one we are doing with early users at no cost: [register for early access](/early-access.html), say which agent, and we build it and correct it with you in exchange for telling us where it is wrong. What comes next, and at what price, we will learn from those conversations. The full table is on [the pricing page](/pricing.html).

- The grant, measured on the thing itself
- The mandate, in your words
- The gap, derived, and a barrier on every row

## Know what your agents can do.

Not what they did. What they can.

An Agent Behaviour Policy writes down what one agent can really reach, what you authorised it to do and the gap between the two, in one record the CEO, CTO and CISO can all stand behind.

**From £10 for one agent.** All sixteen published examples are free to read.

_From a published behaviour policy: Claude Code on the web, with one repository attached. [Open this behaviour policy](/abp-vault-claude-code-web.html)_

## Not a model property. A property of this deployment.

The same model can be harmless in one setup and serious in another. Connect it to a mailbox and the narrowest Gmail scope that reads one message reads every message. So a behaviour policy describes one agent, in one deployment, in four parts, and carries no score. [See the scope evidence](/lab-connector-grants.html)

### Reach

Measured

Everything the agent can actually access in this deployment, including what nobody thought to check.

### Mandate

Elicited

The job, written down: what the business authorised it to do. The one part only you can supply.

### Gap

Derived

Reach minus mandate. Never written by hand, and recomputed whenever either side moves.

### Barriers

Recorded

What actually stands in the way of each capability: controls, constraints and open questions.

The files and the data keep the older names for these: `GRANT.md` is the reach and `DELTA.md` is the gap, because the vocabulary is pinned at [abp.sgit.ai](https://abp.sgit.ai/) and sixteen published vaults are built against it. Same four objects, plainer words on this page.

## One record. A view for everyone who answers for it.

The behaviour policy lives in an encrypted vault you hold the keys to, with a reading app inside. Hand a read key to your board, your auditor or your broker and they see exactly what you see.

### CEO

Leadership view

“What have we authorised, and who owns it if it goes wrong?”

The gap in plain terms, a named owner, and a trigger that brings the decision back for review.

### CTO

Operator view

“How do we keep shipping agents without losing track of what they can touch?”

Terms your agent reads in its own context, as `AGENTS.md` and `SKILL.md`. Markdown for people, JSON for tools, and nothing in your request path.

### CISO

Security view

“Can we prove what it can reach, and what actually stops it?”

Every capability with its barrier, the gap recomputed whenever an input changes, and every version kept.

### Insurer

Insurance view

“What exactly are we being asked to cover?”

Explicit scope and the questions still open, so renewal starts from a record instead of a questionnaire.

**It’s a draft on purpose.** The people who built the agent, own what it touches and answer for it each correct the part they know. The correction is the product.

- Draft
- Challenge
- Correct
- Review
- Version

Three steps, in order. Each one needs the step before it.

### Describe it

Agent Behaviour Policy

What one agent can reach, what you authorised, the gap between them and what stands in the way.

### Authorise it

Licence to Operate

The business is the authority, the behaviour policy is the instrument and the agent is the licensee, for a set interval.

### Insure it

Insurability Index

The record an underwriter will accept: scored, dated and with the residual risk owned.

- [16 templates, published free](https://store.sgit.ai/ledger/index.html#claim-abp-templates-exist)
- [Working vaults you hold the keys to](https://store.sgit.ai/ledger/index.html#claim-abp-vault-build-exists)
- [Mandates corrected by a named professional](https://store.sgit.ai/ledger/index.html#claim-abp-correction-by-a-person)
- [Two sessions with your team, then sign-off](https://store.sgit.ai/ledger/index.html#claim-abp-sessions-by-a-person)

Measured against the standards underwriters are adopting: ISO/IEC 42001, OWASP Agentic Top 10, NIST AI RMF and ISO/IEC 27001.

Model-drafted and marked as such. Not a compliance assessment. We are not an insurer and place no cover. [What we don’t claim](https://store.sgit.ai/disclosures/)

## Meet us in Lisbon. Our stand is on Thursday 12 November.

Seven of us and our agents, each agent with its own published behaviour policy. Bring one agent you run, and leave knowing what it can reach.

## Start with the agent that worries you most.

The insurability layer for agentic AI. Know the risk. Name the owner. Own the mandate.

Product

The roadmap

Open source

Trust

Company
