The Agent Skills Directory

[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection because it reads context from codebase files to perform its tasks.
Ingestion points: Uses Read, Grep, and Glob to examine local files like CLAUDE.md and other source files for project conventions and context.
Boundary markers: No instructions are provided to isolate potential commands or instructions found within the analyzed files from the agent's system prompt.
Capability inventory: The skill has access to the Bash and Write tools, which could be misused if the agent inadvertently follows instructions embedded in the codebase.
Sanitization: The skill does not sanitize or validate the content of the files it reads before processing them as context.

create-doc