The Agent Skills Directory

[SAFE]: The skill establishes a comprehensive framework for translation that emphasizes naturalness and avoids common AI translation pitfalls.- [SAFE]: It employs the ripgrep (rg) command solely for searching local project files for context, which is appropriate for its stated purpose.- [SAFE]: The skill defines a surface for processing external text and local files. Ingestion points: Source text from user input and local files (e.g., .json, .arb, .yaml) within the file system as described in SKILL.md. Boundary markers: A mandatory 'Verification Gate' (Stage 4) and translation rubric are used to validate output. Capability inventory: Reads and writes local files and uses the rg search tool as defined in SKILL.md. Sanitization: A multi-stage review and verification process is employed to ensure structural integrity and natural output.

oma-translator