hyperframes-sales-demo-builder

Pass

Audited by Gen Agent Trust Hub on Aug 28, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data from prospect websites and CRM notes. This introduces a potential surface for indirect prompt injection where malicious instructions embedded in that data could influence the agent's behavior during the demo creation process.
  • Ingestion points: Prospect public sites (via WebFetch/WebSearch), CRM discovery notes.
  • Boundary markers: The skill includes a 'Guardrails recap' section but lacks explicit delimiters for external content.
  • Capability inventory: Uses Bash, Write, and hyperframes-cli for rendering.
  • Sanitization: No specific sanitization or filtering of external data is mentioned.
  • [COMMAND_EXECUTION]: The skill references the use of the Bash tool and a specific command-line utility (hyperframes-cli) to perform the video rendering process. This represents a functional capability required for its primary purpose.
  • [EXTERNAL_DOWNLOADS]: The skill references the use of edge-tts, which is a well-known Python package used for generating neural text-to-speech audio.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 28, 2026, 07:25 AM
Security Audit — agent-trust-hub — hyperframes-sales-demo-builder