The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill uses WebFetch to read content from external competitor websites during the research phase. This creates a surface where a malicious site could attempt to influence the agent's behavior through embedded instructions.
Ingestion points: External websites visited during Phase 2 research, as well as local README.md, package.json, and DESIGN.md files.
Boundary markers: The skill lacks explicit delimiters or instructions to ignore potential commands found within the external content it fetches.
Capability inventory: The skill can execute Bash commands (read/list), Write files, and perform WebSearch and WebFetch operations.
Sanitization: No explicit sanitization or filtering is performed on the data fetched from external websites before it is processed by the agent.
[EXTERNAL_DOWNLOADS]: The skill performs web searches and fetches content from external competitor websites to conduct research. While these operations are necessary for the skill's primary function, they involve interacting with untrusted external sources.
[COMMAND_EXECUTION]: The skill uses Bash to inspect the project structure and read file content (e.g., ls, cat, head). These commands are scoped to typical project files and are used for context gathering.
[DYNAMIC_EXECUTION]: The skill generates a local HTML preview file (design-consultation-preview.html) that includes CSS and JavaScript to demonstrate the proposed design system, including a light/dark mode toggle and a contrast calculator.

design-consultation