The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents methods for ingesting untrusted data from web requests, which establishes a surface for indirect prompt injection.
Ingestion points: SKILL.md describes the use of @Query, @Param, and @Body decorators to extract data from external HTTP requests for use in controller methods.
Boundary markers: The provided documentation and code examples do not specify the use of delimiters or specific instructions for the agent to ignore embedded prompts within the ingested data.
Capability inventory: The framework examples focus on standard web request handling; no high-risk capabilities such as subprocess execution, network exfiltration, or arbitrary file system writes are present in the skill's code snippets.
Sanitization: The skill explicitly documents sanitization and validation techniques using decorators like @Pattern, @Email, and @Size to ensure that data conforms to expected formats and types.

bootgs-validation