The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructions mandate that the agent fetch live documentation from https://alova.js.org to ensure accuracy. This creates a surface for indirect prompt injection where instructions embedded in external content could influence agent behavior. Ingestion points: Documentation URLs listed in SKILL.md (e.g., https://alova.js.org/api/method.md). Boundary markers: Absent. Capability inventory: The agent can perform network requests and data processing via the Alova library. Sanitization: Absent.
[INDIRECT_PROMPT_INJECTION]: The documentation in references/BFF_API_GATEWAY.md contains an example of a transparent reverse proxy implementation using app[method]('*', ...). This pattern forwards incoming requests (including URLs, headers, and bodies) without validation, creating a surface for Server-Side Request Forgery (SSRF) if the target URL is not restricted to an allowlist. This pattern likely triggered the automated reputation scanner alert for malicious content.
[EXTERNAL_DOWNLOADS]: The skill is configured to fetch and process external documentation and configuration guidelines from alova.js.org at runtime.
[COMMAND_EXECUTION]: The skill includes standard documentation for installing library dependencies using the command npm install alova.

alova-server-usage