The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill identifies surfaces where untrusted data is processed to generate the final proposal.
Ingestion points: Defined in SKILL.md, the skill ingests project-specific data from DESIGN.md, feature lists, and other external project sources.
Boundary markers: Absent; the instructions do not specify delimiters or instructions for the agent to ignore potentially malicious embedded content in the input files.
Capability inventory: Capabilities defined across SKILL.md and LAYOUT.md are restricted to generating HTML and CSS for browser-based rendering; there are no network operations, file-system-write capabilities (beyond the generated HTML), or subprocess executions.
Sanitization: Absent; the skill relies on manual verification and "Browser Review" steps described in LAYOUT.md rather than automated sanitization of ingested content.
[DYNAMIC_EXECUTION]: The skill generates code dynamically based on user-provided design tokens and content.
Patterns: As described in SKILL.md and LAYOUT.md, the agent generates standalone HTML/CSS using CSS custom properties for styling. This behavior is the primary intended function of the skill and aligns with standard template-based code generation.

creating-client-proposals