The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data. Ingestion points: The skill inspects the brief and content sources as described in SKILL.md. Boundary markers: No explicit delimiters or instructions to ignore embedded commands are present. Capability inventory: The skill can invoke shell commands (e.g., uv run in references/troubleshooting-common.md) and references other skills with rendering capabilities. Sanitization: No explicit sanitization or validation of external content is documented. This surface is inherent to the slide generation task and falls within the primary intended purpose of the skill.
[COMMAND_EXECUTION]: The references/troubleshooting-common.md file provides a command using uv run to execute a Python script (check_contrast.py) located via an environment variable. This script is used for verifying color contrast ratios, representing a standard operational pattern for inter-skill communication and utility execution within the agent's environment.

create-slide-decks