The Agent Skills Directory

[COMMAND_EXECUTION]: The skill documents the use of pnpm sonamu sync as a core command for synchronizing translation files and updating generated TypeScript types within the project.
[INDIRECT_PROMPT_INJECTION]: The internationalization helpers and locale resolution logic process external data, which defines a vulnerability surface for indirect prompt injection if untrusted data is processed. 1. Ingestion points: Untrusted data enters the agent context through the row object in references/localized-values.md (via localizedColumn) and through function parameters in references/dictionaries.md (via SD). 2. Boundary markers: No explicit delimiters or 'ignore embedded instructions' warnings for the agent are present in the documentation. 3. Capability inventory: The skill documents the ability to execute the pnpm sonamu sync shell command in SKILL.md. 4. Sanitization: The documentation does not mention specific sanitization, escaping, or filtering routines for external content before it is processed by the dictionary functions.

sonamu-i18n