The Agent Skills Directory

[COMMAND_EXECUTION]: The skill performs its core analysis by executing shell commands (specifically ripgrep) and running embedded Python, Ruby, and Awk scripts via the Bash tool. These scripts, found in references/frameworks.md, are used to flatten and compare locale files (JSON, YAML, PO).
[DYNAMIC_EXECUTION]: The skill utilizes dynamic script execution by providing code blocks in its reference documentation that the agent is expected to execute in the environment to process data. This is a functional requirement for the skill's i18n diffing capabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the user's project (locale files and source code) to generate its reports.
Ingestion points: Files like messages/*.json, config/locales/*.yml, and various source templates (.tsx, .vue, .html).
Boundary markers: None specified; the skill does not explicitly instruct the agent to use delimiters for untrusted data in its final report.
Capability inventory: Access to Bash, Write, Edit, and Task tools.
Sanitization: None; the skill copies values verbatim from source files for both reporting and scaffolding.
[EXTERNAL_DOWNLOADS]: The skill suggests using existing environment tools like js-yaml or system-standard ruby -ryaml to avoid introducing new dependencies into the environment, following best practices for tool-based skills.

i18n-check