The Agent Skills Directory

[PROMPT_INJECTION]: Indirect prompt injection surface detected where the skill processes untrusted external data.
Ingestion points: Customer emails and names are extracted from Gorgias tickets in Step 2.
Boundary markers: None present; data is used directly in subsequent prompts and commands.
Sanitization: No explicit sanitization or validation of the extracted email strings is performed before being used in Step 3.
Capability inventory: The skill uses composio execute in Bash to perform searches based on the extracted data.
[COMMAND_EXECUTION]: The skill utilizes shell commands to interact with CRM and support toolkits.
Evidence: Employs composio search and composio execute in Bash to perform data retrieval and cross-referencing operations.
Risk: While the tools are vendor-provided, the interpolation of external data into shell commands creates a potential injection vector if the input data (like email addresses) is maliciously crafted.

contact-sync