The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes the kai-brief CLI tool using arguments directly derived from user input (format, site, and keyword). Specifically, the keyword argument is interpolated into a double-quoted shell command string: kai-brief --format {format} --site {site} --keyword "{keyword}" --output json. This pattern can be susceptible to command injection if the input contains shell metacharacters and the execution environment does not safely handle interpolation.
[PROMPT_INJECTION]: The skill serves as a surface for indirect prompt injection by accepting untrusted input that is then processed and stored for use by downstream skills.
Ingestion points: User-provided keyword, site, and format arguments in SKILL.md.
Boundary markers: No explicit boundary markers or instructions to ignore embedded commands are present in the processing logic.
Capability inventory: Executes shell commands (kai-brief) and uses the Write tool to save files in SKILL.md.
Sanitization: While the format argument undergoes basic validation against a whitelist, the keyword and site inputs are used without apparent sanitization in shell commands and for generating file paths.

content-brief