The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on constructing and executing specific one-liner shell commands to generate random integers based on a randint(min, max) primitive.
Evidence: references/rng-fallbacks.md defines several templates for obtaining random values using node -e, python -c, pwsh -Command, and Unix utilities like shuf, jot, and awk.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied data, such as dice notation (e.g., '3d6+1') or narrative context, which serves as the basis for its randomization commands.
Ingestion points: User-provided prompts specifying dice rolls, list choices, or story questions requiring an 'oracle' result.
Boundary markers: The skill does not employ explicit delimiters for user input, relying instead on the agent's interpretation of the request into numerical ranges.
Capability inventory: The skill utilizes the Bash tool to execute commands through the interpreters allowed in the SKILL.md frontmatter.
Sanitization: Instructions mandate that the agent converts user requests into specific integer-based randint calls, which functions as a validation step before command execution.

dice-roll