The Agent Skills Directory

[SAFE]: The skill is authored by the official vendor and follows industry-standard patterns for SDK integration, including secure handling of API keys and tokens.
[EXTERNAL_DOWNLOADS]: The skill references library metadata and documentation from trusted sources such as Maven Central (repo1.maven.org) and the vendor's official domain (getstream.io).
[COMMAND_EXECUTION]: Uses the first-party getstream CLI for environment setup, token generation, and project discovery. These operations are performed within the scope of intended developer tasks.
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it ingests local project data (Gradle configuration, Android manifests, and Kotlin source code) to perform its functions.
Ingestion points: Reads project-specific files including build.gradle, AndroidManifest.xml, and various Kotlin source files.
Boundary markers: The instructions mandate strict adherence to RULES.md and predefined blueprints to ensure consistent behavior.
Capability inventory: The agent has the ability to read and edit project files and execute CLI commands via the getstream tool.
Sanitization: Relies on standard LLM safety guardrails when processing the content of local project files.

stream-android