linkedin-abm-ad-design

Pass

Audited by Gen Agent Trust Hub on Jul 28, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to perform image processing tasks using the Python Pillow library and suggests the use of headless browsers or HTML-to-PDF tools for generating campaign deliverables. These operations are scoped to the task of asset generation.
  • [EXTERNAL_DOWNLOADS]: The instructions reference a vendor-associated GitHub repository (github.com/ZENABM/linkedin-abm-skills) and the ZenABM platform (app.zenabm.com) for supplemental tools and data integration. These references are consistent with the skill's stated functionality.
  • [DATA_EXFILTRATION]: The skill facilitates the export of campaign data to recognized third-party platforms such as Notion and Google Sheets, and allows for the publication of HTML artifacts to live-artifact tools for user review.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests user-provided strategy documents to drive the generation of campaign content. While this creates a data processing surface, the instructions focus on standard artifact creation and do not involve unsafe interpolation of untrusted data into sensitive control paths.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 28, 2026, 04:39 PM
Security Audit — agent-trust-hub — linkedin-abm-ad-design