The Agent Skills Directory

[SAFE]: The skill uses official HubSpot API endpoints and follows standard authentication practices without any signs of malicious intent.
[COMMAND_EXECUTION]: Employs standard system tools curl and jq to perform network operations and process structured data.
[DATA_EXFILTRATION]: All network requests target api.hubapi.com, which is the official domain for HubSpot's well-known CRM services.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests data from external CRM objects.
Ingestion points: CRM object properties (contacts, deals, companies) fetched from api.hubapi.com in SKILL.md.
Boundary markers: None present in the instructions.
Capability inventory: Shell execution via curl and jq as documented in SKILL.md.
Sanitization: None identified in the provided file.

hubspot