bounce-monitoring-workflow
Pass
Audited by Gen Agent Trust Hub on Sep 29, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local Python scripts (
before.py,execute.py,after.py) to perform administrative tasks within a HubSpot portal. These are executed by the user via theuvtool and represent legitimate functional behavior. - [EXTERNAL_DOWNLOADS]: The documentation references the official
astral-sh/uvGitHub repository for dependency management, which is a well-known and trusted source. - [DATA_EXFILTRATION]: The scripts retrieve workflow and property metadata from the official HubSpot API (
api.hubapi.com). This information is logged locally to CSV files in thedata/audit-logs/directory for auditing purposes and is not transmitted to unauthorized external domains. - [INDIRECT_PROMPT_INJECTION]: The skill processes workflow metadata from the HubSpot API, identifying a data ingestion surface.
- Ingestion points: HubSpot API responses containing workflow names and IDs (referenced in all scripts).
- Boundary markers: Absent.
- Capability inventory: Creation of workflows and contact properties via HubSpot API calls (documented in
execute.py). - Sanitization: Absent. This surface is used for technical inventory and collision checking, which is a standard and expected operation for this management tool.
Audit Metadata