The Agent Skills Directory

[SAFE]: The skill operates entirely locally, producing a single static HTML file without any external network dependencies or remote code execution.
[PROMPT_INJECTION]: The skill contains a vulnerability surface for indirect prompt injection because it ingests untrusted data from a local DESIGN.md file and explicitly permits a section within it ('Agent Prompt Guide') to override its instructions. This is an intended feature of the design workflow rather than a malicious implementation.
Ingestion points: The agent is instructed to read DESIGN.md at the beginning of the workflow in SKILL.md.
Boundary markers: No explicit boundary markers or 'ignore' instructions are provided to separate the design content from the agent's core logic.
Capability inventory: The skill requires file_write to output the final index.html file.
Sanitization: No sanitization or validation is applied to the content retrieved from DESIGN.md.

saas-landing