The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill demonstrates how to process data from external files and APIs, which constitutes an attack surface for instructions embedded in untrusted data sources.
Ingestion points: Instructional snippets using read.csv, read_excel, fromJSON, dbReadTable, and httr::GET in SKILL.md.
Boundary markers: No delimiters or instructions to ignore embedded prompts are included in the code examples.
Capability inventory: The skill is configured to execute R code and scripts using the host's Bash shell via the Bash tool.
Sanitization: No data validation or sanitization logic is shown in the basic import examples.
[DATA_EXFILTRATION]: The skill includes examples of network operations using a reserved documentation domain.
Evidence: The skill demonstrates fetching content from api.example.com using the httr library in SKILL.md.

r-expert