The Agent Skills Directory

[PROMPT_INJECTION]: The skill uses the $ARGUMENTS variable to capture a plugin pack name or domain, which is then used to construct directory structures and file paths.
Ingestion points: The $ARGUMENTS variable in SKILL.md is the primary entry point for untrusted data.
Boundary markers: No specific delimiters or safety warnings are provided to handle cases where the input might contain instructions designed to override agent behavior.
Capability inventory: The skill utilizes Bash, Write, and Edit tools as defined in the YAML frontmatter of SKILL.md.
Sanitization: The instructions do not explicitly mandate input sanitization before using the argument in file system operations or shell commands, although a scripts/validate.py is mentioned as a final checklist step.
[COMMAND_EXECUTION]: The instructions explicitly direct the agent to modify file permissions for generated content.
Evidence: The 'Authoring Rules' section in SKILL.md specifies that optional hooks bundled in a pack must be made executable using the chmod +x command. This represents a controlled but relevant capability for modifying the execution status of local files.

plugin-creator