The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-supplied content and prompts to generate presentations, which presents a surface for indirect prompt injection.
Ingestion points: userContent is ingested in src/gamma/sanitize.ts and src/gamma/exports.ts (SKILL.md).
Boundary markers: The skill does not explicitly define delimiters for user input, though it provides sanitization logic.
Capability inventory: The code snippets include file writing via node:fs/promises and network operations via fetch and @aws-sdk/client-s3 (SKILL.md).
Sanitization: The skill provides a sanitizeContent function that uses regex to mask emails and phone numbers before data is sent to the Gamma API.

gamma-data-handling