Home / Articles / The intersection: SGit, SHarness, Agent Behaviour Policies, RiskMandate and AgentiCCells as one stack of separate companies, now with real customers, real usage and real examples, one ladder of offers on every site, and a plan to align them all
The intersection: SGit, SHarness, Agent Behaviour Policies, RiskMandate and AgentiCCells as one stack of separate companies, now with real customers, real usage and real examples, one ladder of offers on every site, and a plan to align them all
By Dinis Cruz · 2026-11-16 · article v1.1.0, 2 versions · newsroom v0.14.117 · strategysgitsharnessabpriskmandateagenticcellspromptugalvoicedebriefpositioningarticle
Abstract: What we have now is a ridiculously powerful workflow, and it runs across what are really several companies. SGit, encrypted vaults and lanes, is one foundation; SHarness, kernel-enforced Process Behaviour Policies, is another; a key for every person and agent is the third. On top sit Agent Behaviour Policies, the standard that underpins all of it, housed at RiskMandate for now. RiskMandate sells those policies. And AgentiCCells sells agentic teams built on all of it, which is what was missing: RiskMandate had a strong idea and a hard first step, because people asked how to start, and a £500 cell is a way to start. With the site factory, the last manual piece, creating repositories and making DNS and AWS changes, is now a commit, so delivery can scale. Above them are the customers and uses: PrompTugal.news going local town by town, the Web Summit team, the sites the factory makes, five apps, VoiceDebrief, and this newsroom. Money is not changing hands between them, but each uses the others, and that is real usage I can point to. So every site should now sell something on the same ladder of one-off prices, £500, £1,500, £5,000 and £20,000, show its examples, link to the others and to the content here, and say where it sits in the stack. This is the map, and the plan to align them.
In short
- A powerful workflow, across several companies: SGit, SHarness and PKI underneath; Agent Behaviour Policies as the standard; RiskMandate and AgentiCCells on top.
- AgentiCCells is the missing first step: a fixed-price cell, using RiskMandate's policies and SGit's vaults, is how a customer starts.
- The factory removed the last manual piece: repositories, DNS and cloud changes are now a commit.
- Real usage, cross-referenced: PrompTugal.news, the Web Summit team, the factory's sites, five apps, VoiceDebrief, OmKneeHealth and this newsroom each use the layers below.
- One ladder on every site: £500, £1,500, £5,000 and £20,000, each company selling its own thing at each step.
- Next: align every site to the same pattern, and give SHarness its first showcase.
Why now
Until a few weeks ago, these were ideas with demos. Now we have real customers, real usage and real examples. Six of us, with around 250 agents, ran a team operating system, a newsroom, a storm site, a fringe guide and native apps around Web Summit, all on this stack (ten days at Web Summit). PrompTugal.news has gone local, with an agent for every town in the Algarve (every story has an address). And last night the AgentiCCells site factory made a website from a commit in two and a half minutes (the site factory works), which removed one of the last pieces that was still manual and painful.
What that shows is an intersection: a set of companies that each do one thing, and that, together, go end to end. It is time to map it, and to make every site say so.
The foundations
- SGit: git for encrypted vaults. A vault is a unit of work, data, app, history and sources, versioned and handed over with a single read key; lanes in a vault are how our agents message each other with a record. It sits under almost everything here, and its two public examples are the summit and the health scenario with OmKneeHealth.
- SHarness: a Process Behaviour Policy for every process an AI agent starts, what it may read, write, run and connect to, enforced by the Linux kernel. It is the boundary layer: the kind of control that holds because the agent cannot reach it.
- PKI: a key for every person and every agent, made in the browser or the session, with the public half in a registry (a key for every person). It is how we know who did what, and who it was for (identity for agents, layer by layer).
The standard: Agent Behaviour Policies
Agent Behaviour Policies are the underpinning of all of this: for one agent, the grant (what it can reach), the mandate (what it was asked to do), the gap between them, and the barrier that actually stands in the way. The Web Summit team published policies for its agents, PrompTugal.news keeps 45 roles in one policy file, and the same grammar now describes a CI pipeline (CI behaviour policies). For now the standard is housed at RiskMandate; its open home will come later, with the open security community.
RiskMandate, and what it was missing
RiskMandate sells Agent Behaviour Policies for real agents: from a downloaded pack at £10 to a policy corrected for your situation at £500, and two sessions with a professional's signature at £1,500. Its use case is the summit: the team's agents, the stand, and the film (know the reach).
What it was missing was a sale that is easy to start. A policy is valuable, but people asked: how do I start? What do I do on day one? That is the gap AgentiCCells fills.
AgentiCCells: the first step
AgentiCCells sets up small teams of AI agents, in the customer's accounts or ours, as fixed-price cells: one agent or a website at £500, a few sessions handed over at £1,500, a team with a conductor, keys and a board at £5,000, and a network of teams at £20,000 (six setups of an agentic system). Every cell runs on SGit's vaults and lanes, gives each agent a key, and puts each agent under a behaviour policy: AgentiCCells uses RiskMandate, and sells it, as part of something a customer can buy today. With the site factory, delivery is now a commit away (atomic cells, end to end).
The customers and uses
- PrompTugal.news is the customer that uses everything: vault lanes, agent and person keys, behaviour policies for 45 roles, and the AgentiCCells way of running agentic teams, now one agent per town (PrompTugal.news by place).
- The Web Summit team ran on vaults, lanes, keys and published policies, and RiskMandate was its use case (the team's operating system).
- The sites the factory makes, starting with RiskMandate at Web Summit and my brother's site: AgentiCCells customers, in practice if not yet in money.
- Five apps: the PrompTugal news app, the Team app, the RiskMandate app on iPad and an Android tablet, the Lisbon Tech Fringe guide, and VoiceDebrief (native apps by agents).
- VoiceDebrief deserves its own line. With the new iPhone and iPad app, a voice note becomes a transcript and a debrief on the device, kept in a vault the user owns. It is now genuinely usable, and it adds a lot of value: a product of its own, built on SGit.
- OmKneeHealth: Chinmay Gupte, a consultant knee surgeon in London, is building OmKneeHealth, a clinically led platform for knee health, for patients and clinicians he already knows (the expert founders). I have been helping him. Health-related projects use vaults, and his is the public scenario: data that has to stay private, and still be handed over to the right person (where to invest in sgit).
- This newsroom: media and demo vaults, every release signed by a writer key, and more than 420 articles that explain every piece of this. Every site should use them.
Money is not changing hands between these companies, but each uses the others for real, and that is evidence a customer can check.
One ladder on every site
Every site should have something for sale, and the pattern is the same: £500 for one workflow, done in a session; £1,500 for two sessions and a handover; £5,000 for a few days and a team; £20,000 for a bespoke engagement. Each company sells its own thing at each step: a cell, a policy, a vault, a process policy, a set of debriefs, a town page. AgentiCCells and RiskMandate already show prices; the others are proposals in the figure.
Separate companies, one map
I see each of these as a separate company, with its own investment, its own flows and its own go-to-market:
- SGit sells to developers and teams who share data with agents and with each other.
- SHarness sells to platform and security engineers who need a boundary agents cannot cross.
- RiskMandate sells to the people who answer for agents: the CISO, the CTO, the board, and the consultants who advise them.
- AgentiCCells sells to small teams and businesses that want agentic teams set up, seen and stoppable, starting with a website.
- VoiceDebrief sells to professionals and teams who meet, talk and need the record.
- PrompTugal.news serves readers and local businesses, town by town, and shows what all of the above can do.
The plan: align every site
Each site should carry the same few things, in its own words:
- One line on what it is, and where it sits in the stack, with the stack figure and its own layer highlighted.
- The ladder, with what it sells at each step, and a way to buy or start.
- Real examples and customers, with links: the summit, PrompTugal.news, the factory's sites, the apps.
- Links to the others, above and below it in the stack.
- Links to the newsroom's articles that explain it.
- A contact and an outreach path, so a reader who arrives on one site can find the one that fits.
And one gap to close: SHarness has no user in the map yet. It is the foundation with the most to show, and the next showcase should be a team of agents whose processes run under it.
What this does not claim
- No money has changed hands between these companies; the uses are real, the transactions are not.
- Only AgentiCCells and RiskMandate show prices today, and AgentiCCells calls its prices first proposals; the rest of the ladder is a proposal.
- The ticks are only what I could evidence on 16 November 2026; an empty cell is not a no.
- "Separate companies" is how I think about them; their legal structure is not described here.
Where this comes from
My own thinking, 16 November 2026. The live sites: sgit.ai, sharness.ai, abp.sgit.ai, riskmandate.ai, agenticcells.ai, promptugal.news, voicedebrief.ai and its iOS app, read on 16 November 2026; the PrompTugal.news, AgentiCCells and PrompTugal iOS repositories. Earlier on this site: ten days at Web Summit, the site factory works, CI behaviour policies, every story has an address, identity for agents, layer by layer, six setups of an agentic system, atomic cells, end to end, native apps by agents, know the reach.
Peer review Not reviewed yet. Review it
Threads
Builds on
- Ten days at Web Summit: what six people and around 250 agents built between 3 and 13 November, in order, and the series of articles that will tell each part Six people and around 250 agents, ten days at Web Summit: what we built, in order, and the series that will tell it.
- Every story has an address: a design proposal for PrompTugal.news as a ground-up newsroom, where news is filed where it happens and climbs to the region and the country with its proof, mocked up from the live data of 19 local editions A design proposal for PrompTugal.news as a ground-up newsroom: every story filed at an address, climbing with its proof. Mockups from live data.
- The site factory works: three AgentiCCells sites made by hand, then one made by a commit in two minutes and thirty-nine seconds, live over https with a first version written by Claude Code, and the long list of what it took to make something that hard look easy Three AgentiCCells sites by hand, then one by a commit: live over https in 87 seconds, a first version by Claude Code at 159.
- A key for every person: how each person's conductor issues them a key for the team's site, seals what is for them, and how a signed message turns the public web into a channel the agents can trust How each person's conductor issues them a key for the team's site: sealed to you, signed by you, and a lane the agents can trust.
- Identity for AI agents, layer by layer: how our agent teams answer every row of Maya Dotan's map of 190 projects, with a key per agent, a name that says who it acts for, and a behaviour policy that sets its limits How our agent teams answer every row of a map of 190 agent identity projects: names, keys, behaviour policies and a signed record.
- CI behaviour policies: to create one repository and one DNS record per request, our site factory's pipeline holds the power to create, change or delete every repository in the organisation and to point any subdomain anywhere, and why pipelines now need the same behaviour policies as agents Our site factory's pipeline needs one repository and one DNS record per request, yet holds power over the whole organisation. Pipelines need ABPs too.
- Know the Reach: RiskMandate has a song on Spotify, and its music video is being made with AI, shot by shot, from about 2,300 words of human direction; this is how the first part was made RiskMandate's song Know the Reach is on Spotify; the first 53 seconds of its video were made with AI from about 2,300 words of direction.
- Six setups of an agentic system: from a single task to one agent, sub-agents, several sessions, a conductor and its team, and conductors of conductors, with the altitudes of a run inside each and command and control growing at every step Six setups of an agentic system, from a task to conductors of conductors, with the altitudes of a run inside each: make the dangerous thing unreachable.
- Atomic cells, end to end: the £500 website as AgentiCCells' first offering, an order on a queue, an agentic team that builds a public site, a customer who steers it from their browser with a key, and a handover, with no daily work for the founder AgentiCCells' first offering: a £500 website as an atomic cell, steered by the customer from their browser with a key, handed over to their GitHub.
- PrompTugal.news by place: one publication for Portugal, the Algarve and every Algarve town, each place on its own subdomain from the start, each site a lens on one shared semantic graph, and one test for every story, does it connect to AI PrompTugal.news goes local: Portugal, the Algarve and every Algarve town, each on its own subdomain, each a lens on one graph, every story tied to AI.
- The team's operating system: web-summit.sgit.ai, the one site where six people and their agents ran Web Summit, with a brief and a command centre for each person, a key and an inbox for each agent, the board as files, and the outside world on a map web-summit.sgit.ai: one site where six people and their agents ran Web Summit, in four layers built on git.
- Native apps by agents: four iPhone and iPad apps from one workspace and an Android tablet app in five days, from 1,164 lines of Swift to 48,230, a signed Android build four hours after the first commit, and a guide to the fringe on TestFlight ten and a half hours after its brief Agents built four iOS apps from one workspace and an Android tablet app in five days, with TestFlight on every push and a signed build in four hours.
- The expert founders: why the next startups will be built by the people who know the problem, and why Europe's slower, sceptical money suits them Agents make execution cheap, so the people who know the problem can build first. The data favours experts, and Europe's sceptical money suits them.
- Where to invest in sgit: the tool, a hub for vaults, or open core, with the history of developer platforms, the evidence of use, three business plans, and the investor deck as a vault Where investment in sgit should go, from the history of developer platforms, with the deck as a vault.
Continued by
- The cell is the unit: positioning AgentiCCells around small, focused cells, each with an objective, a deliverable and a behaviour policy, each with a page that sells it, a free recipe and three ways to get help, a catalogue that grows one need at a time, in each industry's own words, and why we should run our own work in cells first The cell is the unit of AgentiCCells: one focused job with an objective, a deliverable and a behaviour policy, each with a page that sells it.
- The email agentic cell: why connecting Gmail to agents the usual way is a security mess, and the architecture we will build instead, with one gateway agent that alone touches Gmail, a digital twin of the mailbox in a vault that people and agents work on together, signed proposals, approvals and even reads on append lanes, and behaviour policies at five altitudes Our own email connector: one gateway agent touches Gmail, everyone else works on a signed twin of the mailbox in a vault, with policies at five altitudes.
- Specialist agents for every team: long-lived agents that keep one skill, design, UX, code, security, behaviour policies or writing, with a person who owns it, so that new teams start from our best work instead of from scratch, and why that makes expertise that never scaled into something you can hire, even from another company Specialist agents for design, UX, code, security, ABP and writing serve every team, so new ones start from our best work. Expertise you can hire.
- Integrate with whatever is on the other side: why we never start by asking a customer to move, how we build around the systems they already run, from Odoo 12 to Salesforce, and how, over time, either their system gets better and we retire our code, or ours becomes the better place We never start by asking a customer to move. We integrate with what they run, build around it, and either retire our code or become the better place.
- Odoo's agents can now write to the system of record: how our stack fits around Odoo, from version 12 to 20, with behaviour policies for Odoo agents, an Odoo cell, plan-then-apply through a vault, an ephemeral proof of concept that went from nothing to customised in under a minute on four versions, and why we use whatever version is on the other side Odoo's agents can write to the system of record. Our stack fits every Odoo from 12 to 20: a proof of concept ran on four versions in under a minute each.
Article No. 427 · start a reading list with it
Reviews
No reviews yet.