The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data to generate documentation and code files.
Ingestion points: User-provided parameters such as display_name, scope, and initial_requirements are collected in the Inputs section of SKILL.md.
Boundary markers: The skill does not define specific delimiters or instructions to prevent the agent from obeying malicious instructions embedded within the provided input data.
Capability inventory: The skill has the capability to create new directories and write to multiple files, including .md documentation and .py source files.
Sanitization: While the skill suggests naming conventions (e.g., snake_case), it lacks explicit sanitization or escaping rules for content interpolated into the files.
[DYNAMIC_EXECUTION]: The skill generates Python source code files based on templates and user inputs.
Evidence: Step 7 of the instructions in SKILL.md directs the agent to create a validation.py file with imports and a placeholder structure. The generation of executable source files from potentially untrusted inputs is a documented security surface.

simready-foundation-add-capability