pitch-deck-visuals
Warn
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references an external installation script and documentation hosted at
https://raw.githubusercontent.com/inference-sh/skills/refs/heads/main/cli-install.md. It also instructs the user to install an external package usingnpx skills add belt-sh/cli. - [REMOTE_CODE_EXECUTION]: The skill relies on the
beltCLI tool to perform operations on remote infrastructure (inference.sh). This includes executing Python code, rendering HTML to images, and generating AI images through a remote API. - [DYNAMIC_EXECUTION]: The skill uses
infsh/python-executorto run Python scripts generated at runtime for creating charts (e.g., Market Size, Traction, Competition). This involves dynamically assembling code that is then sent for remote execution. - [INDIRECT_PROMPT_INJECTION]:
- Ingestion points: The skill processes user-supplied data such as company names, problem statements, and financial metrics to populate slide templates in
SKILL.md. - Boundary markers: There are no boundary markers or instructions for the agent to ignore potentially malicious embedded content in the processed startup data.
- Capability inventory: The skill has access to remote Bash execution (via the
belttool), Python execution (infsh/python-executor), and remote image generation tools. - Sanitization: There is no evidence of input validation or sanitization before user-provided text is interpolated into the HTML strings or Python scripts for rendering.
Audit Metadata