The Agent Skills Directory

[Indirect Prompt Injection] (LOW): The skill is susceptible to indirect prompt injection due to the ingestion of untrusted external data.
Ingestion points: Data returned from DOCUSIGN_LIST_ALL_TEMPLATES, DOCUSIGN_GET_TEMPLATE, and DOCUSIGN_GET_ENVELOPE (e.g., template names, descriptions, or envelope content).
Boundary markers: None specified in the instructions to separate external data from agent instructions.
Capability inventory: The agent can send emails and create documents using DOCUSIGN_SEND_ENVELOPE and DOCUSIGN_CREATE_ENVELOPE_FROM_TEMPLATE.
Sanitization: No mention of sanitization or validation of the data retrieved from DocuSign before processing.
[Remote Code Execution] (SAFE): The skill references a remote MCP endpoint (https://rube.app/mcp), which is a configuration for the Rube framework. No direct execution of untrusted scripts (e.g., curl | bash) was detected.
[Data Exfiltration] (SAFE): No evidence of unauthorized data transfer or hardcoded credentials. Authentication is handled via a managed OAuth connection tool (RUBE_MANAGE_CONNECTIONS).
[Obfuscation] (SAFE): No instances of Base64, zero-width characters, or other encoding techniques intended to hide malicious behavior were found.

docusign-automation