The Agent Skills Directory

[SAFE]: The skill provides structured guidance for R package preparation, focusing on documentation standards, metadata verification, and URL validation.
[SAFE]: All tools mentioned, such as usethis, devtools, and urlchecker, are widely recognized and official utilities within the R development ecosystem.
[SAFE]: External URLs referenced in the skill are restricted to illustrative examples or official R infrastructure, including CRAN, r-pkg.org, and R-Universe.
[INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading and analyzing local package files such as DESCRIPTION, README.md, and R source files. While these files are external data that could contain malicious instructions, the risk is inherent to the code-review functionality of the skill.
Ingestion points: The skill reads DESCRIPTION, README.md, README.Rmd, R/*.R, LICENSE, NEWS.md, cran-comments.md, and .Rbuildignore as part of the review process.
Boundary markers: None explicitly defined to separate user-provided file content from system instructions.
Capability inventory: The skill proposes file edits and suggests the execution of standard R development commands for project maintenance.
Sanitization: No explicit sanitization or escaping of the content within the reviewed files is mentioned.

cran-extrachecks