The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The instructions require the agent to ingest and analyze existing project files (neighbouring DTOs, mappers, and repositories) to identify conventions such as serialization and result handling.
Ingestion points: SKILL.md (Step 1) and references/current-guidance.md (Detect The Repo's Conventions First section).
Boundary markers: The skill does not define specific delimiters or instructions to ignore potential malicious content within the analyzed repository files.
Capability inventory: The agent is authorized to define data persistence logic (Jetpack DataStore, Room) and network service interfaces.
Sanitization: No explicit rules are provided for sanitizing or validating the patterns extracted from the existing codebase.
[EXTERNAL_DOWNLOADS]: The skill references official Android developer documentation and codelabs hosted on Google's domains (developer.android.com). These references are for educational purposes regarding Jetpack DataStore implementation.
[SAFE]: The skill includes numerous security best practices, such as prohibiting the reuse of Response DTOs as Request bodies, preventing the leakage of sensitive transport models to the UI, and enforcing strict nullability checks to prevent data corruption from malformed server responses.

android-state-data