Evaluate extensions to the capability-declaration vocabulary

Determine whether extending the Agent Skills frontmatter vocabulary to express credential and environment access changes what authors declare, rather than merely expanding what they could declare.

Background

The paper finds that the Agent Skills format cannot declare credential or environment access, while proposals to add these fields remain open in the standard’s repository. The authors identify the current adoption and declaration-accuracy measurements as a baseline against which any future vocabulary extension could be assessed.

The unresolved issue is behavioral and empirical: adding fields may increase accurate disclosure, or authors may simply continue declaring at similar rates. The paper does not evaluate an extended vocabulary because the relevant fields do not yet exist.

References

A vocabulary that can express what matters. The format cannot declare credential or environment access, and proposals to add exactly those fields are open in the standard's own repository. If any of them ships, the uptake and accuracy measured here become a baseline. Whether extending the vocabulary changes what authors declare, or only what they could declare, is an empirical question that cannot be asked until the field exists.

— Beyond Maliciousness: Static and Runtime Attack-Surface Profiling of Agent Skills  (2610.19264 - Pandey et al., 13 Oct 2026) in Section Future Work, paragraph “A vocabulary that can express what matters.”