The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data which could potentially harbor malicious instructions.
Ingestion points: Functions such as createAndSendDocument (SKILL.md) and create_document (implementation-guide.md) accept external file paths and user-provided signer information.
Boundary markers: There are no explicit instructions or delimiters provided to prevent the agent from obeying instructions that might be embedded within processed PDF files or metadata.
Capability inventory: The skill utilizes file system access (fs.readFileSync, Python open) and network communication (Documenso SDK, urllib.request, fetch) to handle and transmit data.
Sanitization: The implementation guide includes Zod-based validation patterns to ensure data integrity for API responses and inputs.
[EXTERNAL_DOWNLOADS]: The skill references documentation and source code from legitimate external sources related to the integration.
References the official Documenso GitHub organization for SDK source code and documentation.
Provides links to the author's infrastructure and developer resources at jeremylongshore.com and intentsolutions.io.

documenso-sdk-patterns