# kaal:position:2026-09-12-002

**Affirmed position.** Cao leaves constraints to human articulation. Kaal's position is that authority must be structural instead. The agent proposes, a separate component that alone holds the credentials decides, and irreversible actions go to a named person. Reputation may inform discovery, routing, pricing, and allocation, but it may never grant permission, enlarge authority, or substitute for an enforcement boundary.

**Status.** affirmed  **Published.** 2026-09-12

**Holds when.**

- Cao's paper is a position paper citing benchmark evidence, not an empirical test of these claims.
- The position is Kaal's own argument in the essay Prompts Don't Govern Agents. Reputation Does.
- The position does not validate Cao's figures, benchmark methodology, or reported benchmark results.
- The position is limited to credentialed actions and irreversible agent effects.

**Current debate.** Zhenfeng Cao, Agentic Software: How AI Agents Are Restructuring the Software Paradigm (arXiv:2606.05608v2, 2026): https://arxiv.org/abs/2606.05608

**Extends.** kaal:claim:7314479-020: https://wulfkaal.github.io/claims/7314479-020

**Scholarly basis.** Wulf A. Kaal, Institutional Requirements for Sovereign Local Agent Runtimes (2026). SSRN: https://ssrn.com/abstract=7314479

**Source PDF sha256.** `debace24a155ae924a155b1fafe98856d98cf83689feff2f87a32f1c06171ce6`

**Topics.** reputation, consensus-and-security, governance-design

**Provenance.** Affirmed in kaalvis:2026-09-12:prompts-reputation-essay-six-v1 at https://wulfkaal.com/2026/09/12/prompts-dont-govern-agents-reputation-does/.

**Record type.** This is a dated commentary position that extends a scholarly corpus claim. It is not a verbatim claim extracted from the paper.

**Canonical form.** This markdown file is the canonical hashed representation of the position.