# kaal:claim:7314479-024

**Claim.** Evaluation must derive from observed results under stated conditions, be specific to context, and be costly to manipulate, which excludes self-declared capability, averaged scores across incommensurable tasks, and most engagement metrics.

**Type.** normative  **Support.** argued

**Holds when.**

- coordination layers for sovereign local agent runtimes

**Source quote.**

> Evaluation must derive from observed results under stated conditions, must be specific to context, and must be costly to manipulate. Each clause excludes a common practice. Observed results excludes self-declared capability. Stated conditions excludes aggregate scores that average across incommensurable tasks. Costly to manipulate excludes any signal a participant can generate at will, which in practice excludes most engagement metrics.

**From.** Wulf A. Kaal, *Institutional Requirements for Sovereign Local Agent Runtimes* (2026), VI. Six Institutional Requirements, Requirement 4, page 12

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

**Verify.** sha256 of source PDF `debace24a155ae924a155b1fafe98856d98cf83689feff2f87a32f1c06171ce6` at https://raw.githubusercontent.com/wulfkaal/Academic-Papers/main/papers/pdf/Kaal%20-%202026%20-%20Institutional%20Requirements%20for%20Sovereign%20Local%20Agent%20Runtimes.pdf

**Topics.** reputation, research-methods, risk-and-incentives

**Canonical form.** This markdown file is the canonical hashed representation of the claim. Its sha256 is the content hash used for attestation.