The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes list names, filter criteria, and membership metadata retrieved from the HubSpot API. This data is untrusted as it originates from an external system and could theoretically be manipulated to contain instructions for the agent.\n
Ingestion points: Stage 2 in SKILL.md uses a Python script to fetch list data from https://api.hubapi.com/crm/v3/lists/search.\n
Boundary markers: The instructions do not define explicit delimiters or warnings for the agent to ignore embedded instructions within the HubSpot data.\n
Capability inventory: The skill uses Python scripts to perform network requests (requests) and instructs the agent to delete resources (DELETE via API).\n
Sanitization: There is no explicit evidence of sanitization or validation of the list names or criteria before they are processed or presented for deletion.\n- [COMMAND_EXECUTION]: The skill includes Python scripts that are intended to be executed in the agent's environment to search for and iterate through CRM records.\n- [EXTERNAL_DOWNLOADS]: The skill references the use of the uv package manager and standard Python libraries such as requests and python-dotenv for API communication and environment management.

cleanup-lists