The Agent Skills Directory

[Indirect Prompt Injection] (HIGH): The skill is designed to process external untrusted data from a project directory. * Ingestion points: All files within the provided <project_path>. * Boundary markers: Absent; no instructions are provided to the agent to treat the file content as untrusted or to delimit the data from instructions. * Capability inventory: Execution of a Python script on the host system via the 'Script' section. * Sanitization: Unverifiable as the script source code is not provided.
[Unverifiable Dependencies] (HIGH): The skill relies on an external script, scripts/i18n_checker.py, which is missing from the analysis bundle. Executing unvetted scripts on local file systems is a high-risk operation that could lead to unauthorized file access, data exposure, or system compromise. The use of <project_path> as an argument suggests the script will crawl and process an unknown number of files, increasing the attack surface.

i18n-localization