The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill includes patterns for bulk ticket creation by reading from external data sources like support_requests.jsonl. This represents a potential attack surface where untrusted content could be interpolated into CLI commands without sufficient sanitization.
Ingestion points: The skill reads external JSONL data from support_requests.jsonl to populate CRM fields (Step 3).
Boundary markers: No explicit delimiters or boundary instructions are used to separate user-provided ticket data from the agent's internal logic.
Capability inventory: The skill uses hubspot objects create, hubspot objects update, and hubspot associations create to modify data in a HubSpot portal.
Sanitization: There is no evidence of data sanitization or validation performed on the ingested ticket content before it is processed.
[COMMAND_EXECUTION]: The skill relies on executing the hubspot CLI tool and shell utilities such as jq and cat to perform ticket management tasks. This behavior is aligned with the skill's stated purpose of interacting with the HubSpot API through the command line.

ticket-resolution