The Agent Skills Directory

[SAFE]: No security issues detected. The skill uses standard research orchestration patterns and documented prompt engineering techniques (such as expert personas and stakes language) to improve output quality.
[DATA_EXPOSURE]: The skill is designed to analyze codebases and documentation, which involves reading local files. This behavior is transparently documented and consistent with its stated purpose of "comprehensive research".
[EXTERNAL_DOWNLOADS]: The skill mentions a web-search-researcher agent for gathering external context, which is a standard capability for research-oriented tools.
[PROMPT_INJECTION]: The skill utilizes "Stakes Language" and "Expert Personas" as part of an incentive-based prompting strategy. These are intentional design choices used to focus the model's reasoning and do not represent attempts to bypass safety filters or override system instructions.

comprehensive-research