The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a framework for managing file system access based on user-provided paths and instructions. Since it facilitates the processing of external path strings and boundary definitions, it is theoretically susceptible to indirect prompt injection if an attacker-controlled input (e.g., a file name or path read from a repository) contains instructions designed to bypass the 'freeze' or 'guard' rules. Severity is LOW as it relies on standard agent instruction-following guardrails.- [NO_CODE]: The skill consists entirely of Markdown instructions and metadata without any attached scripts, executables, or network-requesting code.

scope-guard