The Agent Skills Directory

[SAFE]: The helper script scripts/crawlora.sh implements robust security measures, including validation of the API key format, secure header handling via temporary files with restricted permissions to prevent leakage in process lists, and a strict whitelist of allowed API paths and methods.
[INDIRECT_PROMPT_INJECTION]: As a research tool, the skill ingests external game descriptions and reviews, creating a potential surface for indirect prompt injection. This is handled as an inherent characteristic of the skill's primary function. 1. Ingestion points: Data is retrieved from the /xbox/search, /xbox/game, and /xbox/reviews endpoints. 2. Boundary markers: No explicit delimiters are specified in the instructions for external content. 3. Capability inventory: Capabilities are strictly limited to performing GET requests to the vendor's API. 4. Sanitization: Content is processed without explicit sanitization, which is standard for catalog research tasks.

xbox-research