The Agent Skills Directory

[SAFE]: The skill operates as a purely instructional framework for orchestrating other research-related skills (e.g., exa-search, deep-research). It does not contain executable scripts, external package dependencies, or direct network-active commands.
[SAFE]: The instructions explicitly mandate the separation and labeling of user-supplied context from sourced facts and inferences. This implementation of evidence boundaries is a security best practice that mitigates risks associated with data confusion and indirect prompt injection during research tasks.

research-ops