The Agent Skills Directory

SAFE (SAFE): The skill is composed entirely of Markdown documentation and templates. There are no executable scripts (.py, .js, .sh) or configuration files that trigger automated actions.
Indirect Prompt Injection (LOW):
Ingestion points: The skill facilitates the creation of new files based on user-provided strings (name, description, functionality). This creates a surface where a user could provide malicious instructions to be embedded into the generated files.
Boundary markers: The templates use standard YAML frontmatter delimiters (---) and Markdown headers to separate metadata from instructions.
Capability inventory: The primary capability is the generation of text files (SKILL.md, README.md) on the local filesystem.
Sanitization: The skill relies on the base AI model's inherent safety filters and does not include custom sanitization logic for the templates, which is appropriate for its purpose as a template library.
External Downloads (SAFE): While the templates (e.g., 06_utility_tool.md) provide examples of network operations like curl or npm install, these are illustrative placeholders within documentation and do not represent active execution patterns within the skill itself.
Credentials (SAFE): No hardcoded API keys or secrets were detected. The templates specifically advise users to use environment variables for sensitive data.

skill-creator