The Agent Skills Directory

Indirect Prompt Injection (LOW): The skill processes external content stored in the docs/ directory. Since this content is fetched from source_url, it could contain malicious instructions (Indirect Prompt Injection) designed to hijack the agent's behavior when retrieved. No specific boundary markers or sanitization logic are defined in the instruction set.
Command Execution (LOW): The skill invokes a local script scripts/search_docs.py via a shell command with a user-provided <query>. This creates a standard attack surface for command injection if the script or the shell wrapper does not correctly sanitize the input query.
No Code (INFO): The functional script scripts/search_docs.py and the docs/ content were not provided for review. This analysis is based on the usage patterns described in the SKILL.md file.

tailwind