The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user or agent to install external packages using npx skills add belt-sh/cli and npx skills add inference-sh/skills. These packages originate from third-party sources not included in the primary trusted vendor list.
[COMMAND_EXECUTION]: The skill relies on the belt CLI tool and the Bash tool to execute commands such as belt login and belt app run. This allows for local command execution and interaction with the inference.sh remote API.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection where untrusted data could be interpolated into the HTML templates used for image generation.
Ingestion points: The html field in the --input JSON for belt app run infsh/html-to-image (found in SKILL.md).
Boundary markers: None identified; there are no explicit instructions to the agent to treat input as data rather than instructions.
Capability inventory: The skill uses Bash(belt *) to execute shell commands and perform network operations via the belt CLI.
Sanitization: No evidence of input sanitization or validation before passing content to the html-to-image tool.

og-image-design