The Agent Skills Directory

[SAFE]: The skill is a formatting template that generates ASCII and JSON design specifications. No patterns related to data exfiltration, prompt injection, or remote code execution were found.
[INDIRECT_PROMPT_INJECTION]: The skill defines an input model for processing external data (such as error messages and component properties).
Ingestion points: The Input Model in SKILL.md accepts external values via props and state.error fields.
Boundary markers: Output is strictly encapsulated within specific labeled blocks like TUI_RENDER and COMPONENT_SPEC.
Capability inventory: The skill does not request or use any tools for file writing, network requests, or shell command execution.
Sanitization: The rendering rules specify that error content must be truncated to a single line footer, limiting the impact of malicious input.
Conclusion: The ingestion surface is safe due to the lack of exploitable capabilities.

tui-count-to