The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill functions as a prompt generator that ingests user-provided strings (repository identifiers, labels, CI check names, and notification destinations) and interpolates them into a template for future execution by an agent.
Ingestion points: Data is collected from the user via interactive prompts to fill placeholders like {owner/repo}, {名前}, {checkの名前}, and {通知先} in the SKILL.md file.
Boundary markers: The resulting prompt template does not include delimiters or instructions to prevent the agent from interpreting instructions that might be embedded within the user-provided values.
Capability inventory: The skill leverages platform-specific capabilities to list, create, and update 'routines', which are scheduled execution environments for the agent.
Sanitization: There is no evidence of input validation or escaping for the user-provided data before it is placed into the prompt template.
[PERSISTENCE]: The skill's primary objective is to establish 'routines', which are persistent scheduled tasks that run automatically. While creating persistence mechanisms is often a security concern, it is the core functionality of this specific setup tool and is explicitly requested and configured by the user.

dependency-update-setup-routine