Declared Purpose & Out-of-Scope Register
Every agent has a written, versioned statement of its intended purpose, the user populations it serves, and the actions it must never take.
This page is the complete instruction page for one pattern called "Declared Purpose & Out-of-Scope Register." It explains the problem this pattern solves, the idea behind the solution, when you should (and should not) use it, and what happens afterward — both the good effects and the costs.
This matters because building AI agents is not just about making them clever. It is also about making them safe and predictable. Following a well-tested pattern like this one helps avoid common mistakes, and shows you exactly which safety rules and regulations it connects to, listed under "Standards Mesh" on this page.
Context
Agents are often deployed with purpose stated only implicitly in a system prompt. Over time, capability creep produces agents doing things no one authorised — "feature drift" that regulators classify as unmanaged risk.
Problem
Without a ground truth for purpose, no reviewer can tell whether a behaviour is intended, emergent, or broken.
Forces
- Generality of LLMs vs specificity of governance
- Documentation overhead vs audit readiness
Solution
Maintain a short charter (10–20 pages) per agent stating: system description, intended purpose, user populations, deployment environment, prohibited uses (enumerated, not hand-wavy), change-review process. Prohibited uses must be enforceable by the guardrail layer, not merely stated in prose.
Applicability
- All production agents
- Any agent subject to regulatory review
- Enterprise agent portfolios
Anti-Patterns
- Purpose stated only in code comments
- 50-page boilerplate no one reads
Consequences
- +Stable reference for audits and incident reviews
- +Bounds feature-creep
- −Requires maintenance discipline