The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The CI workflow in references/implementation.md fetches the Lokalise CLI installation script from Lokalise's official GitHub repository.
[COMMAND_EXECUTION]: The skill utilizes shell commands for translation promotion, environment setup, and secret management using tools like jq, lokalise2, and CLI interfaces for AWS, GCP, and HashiCorp Vault.
[INDIRECT_PROMPT_INJECTION]: The skill processes translation data from external Lokalise projects, creating a potential attack surface for indirect prompt injection.
Ingestion points: The lokalise2 file download command in SKILL.md and references/implementation.md fetches external JSON translation files into the local environment.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potentially malicious content within the fetched JSON files.
Capability inventory: The skill environment permits Bash execution (specifically for aws, gcloud, and vault namespaces), along with Read and Write file system access.
Sanitization: The skill does not perform sanitization or validation of the content within translation files before processing them with jq or re-uploading them to different environments.

lokalise-multi-env-setup