The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute local maintenance and audit scripts (node scripts/check-translation.mjs, node scripts/audit-english.mjs) and use Git commands (git diff) to verify repository integrity and check for synchronization issues. These are standard operations within the context of managing a software repository.
[INDIRECT_PROMPT_INJECTION]: The skill is intended to process external data (content from the upstream mattpocock/skills repository). This represents an attack surface for indirect prompt injection if the upstream content contains malicious instructions. However, the skill provides extensive instructions on distinguishing between translatable prose and behavior-critical identifiers, and explicitly directs the agent to protect the latter, which serves as a mitigation.
Ingestion points: Content files from the mattpocock/skills repository.
Boundary markers: The skill defines explicit categories for "What to translate" vs "What to preserve exactly" and includes instructions to "protect behavior-critical spans before translation".
Capability inventory: The agent is authorized to run local scripts using node and perform Git operations.
Sanitization: The agent is instructed to use a fail-closed rule: if unsure whether text is behavior-critical, it must be preserved and flagged for review.

translate-skill