method candidate goes to: this-edition V villager
Every rule lands in one of three states: backed by a control, governed by an accepted risk, or resting on hope
For each rule the policy records what stands in the way: nothing, an expectation (a rule in prose, enforced by nobody), a setting the agent's own account can flip, or a boundary enforced above it, and only the boundary is a control. Add one more fact, whether a named person has accepted the risk of the rule not being enforced, with an expiry, and every rule is a control, an accepted risk or hope, which can be counted. In the article's fictional firm 5 of 17 rules are backed by a control, 6 by an accepted risk and 6 by hope; the platform layer is all control, the organisation layer all hope, and each existing product (a send gateway, an outbound content check, a blocked-recipients list) moves one row.
Decided by. Nobody yet: a candidate for Dinis to read, comment on and decide (keep, rewrite, merge, cut, article first, later, not doing it).
Note. Adds the third state, accepted risk, to mandate-reach-gap's hope versus enforcement. The proportions are illustrative; the article says the shape is the point. Resting on hope is not a failure, most of a business has always run on people behaving as asked, but with an agent it can be counted.
Where it comes from
- Zoom into an agent's behaviour policy and you find the business logic How much of it is hope?; Where the vendors plug in (newsroom)
What it answers in the 2018 book
- Threat Modeling answered outline, Security
Where it goes
Chapter Controlling your agent: RiskMandate, part Build it without a server (drafting); topic Controlling your agent: mandate, reach and hope versus enforcement.
Other ideas proposed for the same chapter: Mandate, reach and the gap: hope versus enforcement, Code is a liability; the future is the smallest code that does the job, Learning to hack is still how you learn how things work: draw the line where the attacker is, The secret goes from the drop into the tool, never into the chat, Security experience is what lets you ship with known gaps: list them, match them to the reader, say what would change the call, The agent behaviour policy is the due diligence document: a buyer reads the blast radius off it, and its absence says the most, A vault per conversation and per job: a leaked key exposes one small thing, not an inbox, The desktop is the barrier, not the model: prompt injection is not solved, so the box has to hold, Reach is a computed value, and new capability lands as excess, never as new authority, Who can pull the plug: an agent that cannot be stopped is riskier than one that can, A third input: a behaviour policy beside the system prompt and the prompt, which the vendor acknowledges, enforces where it can and answers for, Models do not separate code from data, so a rule in the prompt can be argued with by an email, Aviation and payments got safe when the rule, the liability and the business model pointed the same way, Zoom into an agent's behaviour policy and you find the business logic, which only the business can write, Software was the law: the rule the screen enforced by omission is no longer enforced once an agent calls the API, Write down what your agent can reach in twenty minutes: list its tools, say what you wanted it for, write the gap.
Node idea:control-accepted-risk-or-hope in graph.json; see it in the graph. Source: admin/content/graph/ideas.json.