analytics-dashboard

Pass

Audited by Gen Agent Trust Hub on Apr 28, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill uses the 'CRITICAL' marker to instruct the agent to override its default interaction flow and immediately begin the analytics workflow upon loading.
  • [INDIRECT_PROMPT_INJECTION]: The skill parses data from an external, potentially attacker-controlled Excel file without safety boundaries.
  • Ingestion points: Processes multiple sheets from a user-uploaded LinkedIn Analytics .xlsx file in Step 2.
  • Boundary markers: Absent; the skill does not use delimiters or instructions to ignore embedded commands within the spreadsheet data.
  • Capability inventory: Generates React code (artifacts) and written strategic recommendations based on the data.
  • Sanitization: Absent; the skill performs header cleaning for formatting but does not sanitize cell content for malicious instructions.
  • [DYNAMIC_EXECUTION]: The skill dynamically generates a React artifact using the Recharts library, incorporating data extracted from the external file into the rendered code.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 28, 2026, 07:05 PM