The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/install-linguist-hooks.sh installs a Git hook in the .git/hooks directory. While this is a standard feature for automating development workflows, it functions as a persistence mechanism within the repository's configuration environment.
[COMMAND_EXECUTION]: The script scripts/install-linguist-hooks.sh contains a path traversal vulnerability. The --mode parameter is used to construct a file path (.git/hooks/${MODE}) without sanitization, allowing files to be written to arbitrary locations if the script is invoked with malicious path segments.
[COMMAND_EXECUTION]: The skill's documentation and the PowerShell wrapper scripts/linguist-i18n.ps1 utilize the -ExecutionPolicy Bypass flag. This facilitates script execution by bypassing local security restrictions on Windows systems.
[PROMPT_INJECTION]: The skill processes codebase files through scripts/check-linguist.mjs, creating a surface for indirect prompt injection.
Ingestion points: The script reads and processes project files (JSON, JS, HTML, etc.) in scripts/check-linguist.mjs to detect hardcoded strings.
Boundary markers: No delimiters or specific instructions are used to isolate untrusted project content from the agent's logic.
Capability inventory: The skill can write files and install Git hooks through its included scripts.
Sanitization: No sanitization or filtering is performed on the content of the analyzed files before processing.

linguist