# kaal:position:2026-08-28-005

**Affirmed position.** Mui, Mohtashemi, and Halberstadt provide a formal basis for Kaal's domain-specificity commitment. Their model indexes an agent's reputation by context and by the evaluating agent, and encounter history is accumulated within a context rather than globally. They state that cross-context estimation is not addressed and identify vectorized quantities as a simple design direction. This extends Kaal's non-transfer rule: without an explicit cross-context estimator, silent domain transfer is not licensed by the model. The source does not test silent-transfer harms, measure how often implementations discard context, or attribute context collapse to display convenience. The reviewed evidence is the authors' PostScript copy, not publisher PDF bytes.

**Status.** affirmed  **Published.** 2026-08-28

**Holds when.**

- Authors' PostScript evidence rather than publisher PDF bytes.
- Cross-context estimation and silent transfer are not evaluated.
- No evidence on implementation prevalence or display convenience.

**Current debate.** A Computational Model of Trust and Reputation: https://groups.csail.mit.edu/medg/people/lmui/papers.html

**Extends.** kaal:claim:7314479-045: https://wulfkaal.github.io/claims/7314479-045

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

**Source PDF sha256.** `debace24a155ae924a155b1fafe98856d98cf83689feff2f87a32f1c06171ce6`

**Evidence level.** permitted_full_text

**Mapping review tier.** independent evidence-first compound-growth review

**Mapping confidence.** unscored  **Mapping ambiguous.** false

**Topics.** reputation, governance-design, decentralization, ai-and-agents

**Provenance.** Affirmed in kaal-growth:2026-08-28:daily-001-remediated-v1 at https://wulfkaal.github.io/positions-src/2026-08-28-compound-growth-kaal-growth-2026-08-28-daily-001-remediated-v1.json.

**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.