The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill invokes the infsh/python-executor application to run dynamically generated Python code. While the examples provided are for Matplotlib visualizations, this pattern allows for the execution of arbitrary code within the agent's operating environment.
[EXTERNAL_DOWNLOADS]: The skill instructs the user to install external tools via npx (belt-sh/cli) and fetch installation guidelines from a remote GitHub repository (inference-sh/skills). These resources originate from sources outside the pre-approved trusted vendor list.
[COMMAND_EXECUTION]: The skill requires access to the Bash tool to interact with the belt CLI. This includes performing authentication (belt login) and running remote applications, which increases the complexity of the agent's interaction with the host system.
[INDIRECT_PROMPT_INJECTION]: The skill provides a surface for indirect injection as it ingests user-provided pitch deck data into shell command arguments (--input JSON).
Ingestion points: Data is passed into the html and code fields of the belt app run command in SKILL.md.
Boundary markers: None present; the JSON strings are directly interpolated.
Capability inventory: The skill has access to Bash via the belt CLI and can execute Python code.
Sanitization: There is no evidence of input escaping or validation for the provided data structures.

pitch-deck-visuals