The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions to run npx --yes submit-expo-feedback@latest and npx testflight. Both commands download and execute code from the npm registry at runtime. Neither package is associated with a trusted organization or the verified namespace of the skill author. The command npx testflight is particularly concerning as it is presented as a shortcut for official Expo tasks but targets a third-party package not recognized as part of the official Expo toolchain.
[CREDENTIALS_UNSAFE]: The skill requires the management of sensitive secrets, including Apple distribution certificates (.p8 files), Google service account JSON keys, and EAS environment secrets. These files contain credentials that provide high-level access to production app store accounts and CI/CD pipelines.
[INDIRECT_PROMPT_INJECTION]: The skill processes external data that could be leveraged for prompt injection attacks, particularly through app store metadata and external API fetches.
Ingestion points: External metadata pulled via eas metadata:pull and translations fetched from remote APIs in references/app-store-metadata.md.
Boundary markers: Absent. The instructions do not provide delimiters or warnings to treat external data as untrusted.
Capability inventory: The skill utilizes eas-cli and npx (as referenced in SKILL.md), which possess capabilities for filesystem modification, network requests, and shell command execution.
Sanitization: Absent. There are no mechanisms described for validating or escaping content retrieved from external sources before it is processed by the agent.

expo-deployment-v55