The Agent Skills Directory

[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it is designed to ingest and process untrusted data from users and other agents to generate slide content.
Ingestion points: User-provided outlines and specifications, as well as structured handoffs from other agents like 'Scribe' (defined in references/handoffs.md).
Boundary markers: The instructions lack explicit delimiters or instructions for the agent to ignore or isolate potentially malicious instructions embedded within the source content.
Capability inventory: The skill has the capability to generate executable shell commands and file content based on the input provided.
Sanitization: There is no evidence of input validation or sanitization before the content is interpolated into generated slides or commands.
[EXTERNAL_DOWNLOADS]: The skill references well-known technology tools and official package registries for its core functionality.
Evidence: Mentions @marp-team/marp-cli and the hono framework in SKILL.md and references/examples.md for project initialization and slide previewing.
[COMMAND_EXECUTION]: The skill provides the user with shell commands necessary to use the supported frameworks.
Evidence: Includes commands such as npx @marp-team/marp-cli slide.md --preview and npm create hono@latest my-api in the documentation and examples.

stage