The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill is designed to generate new agent skill files (SKILL.md) at runtime by scaffolding them from templates. This allows for the creation of new executable agent logic within the project's environment.
[DATA_EXPOSURE_AND_EXFILTRATION]: The skill accesses the user's home directory to read template files located at ~/.claude/skills/_templates/feature/. While these are expected configuration paths for this platform, the skill requires access to sensitive directory structures outside the immediate project workspace.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided input (the feature name) to create file slugs and populate content.
Ingestion points: The <name> argument provided via the /feature-define trigger.
Boundary markers: No explicit boundary markers or 'ignore' instructions for the interpolated data are defined in the provided file.
Capability inventory: The skill has the capability to write to the filesystem (.claude/skills/, ai-context/features/) and modify project configuration files (CLAUDE.md).
Sanitization: The instruction does not specify validation or sanitization routines for the user-supplied name before it is used to generate filenames or content.

feature-define