The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted text from local project files during analysis, creating a surface for indirect prompt injection if those files contain malicious instructions meant to influence the agent.
Ingestion points: The scripts/i18n_checker.py script reads the content of code files (.tsx, .jsx, .vue, .py) and translation files (.json, .po) to identify untranslated strings.
Boundary markers: Absent. The script identifies and prints hardcoded strings directly to the agent's output buffer without using delimiters or instructions to treat the content as untrusted data.
Capability inventory: The skill is limited to reading local files and printing to standard output; it does not possess inherent network, file-write, or shell execution capabilities beyond the checker script itself.
Sanitization: Absent. The script extracts strings using regular expressions and displays them verbatim without filtering for potentially malicious commands or prompt injection patterns.

i18n-localization