The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill requires the agent to read and interpret external static HTML mockups and CSS token sources, which represents a surface for indirect prompt injection if the design files are compromised. Evidence: 1. Ingestion points: The agent is instructed to read 'mockup HTML' and 'CSS token source' files (SKILL.md). 2. Boundary markers: The process includes 'Gate 1' and 'Gate 2' which mandate operator approval before implementation and finalization (SKILL.md). 3. Capability inventory: The agent has capabilities to modify Twig/SDC files, execute git commands, and run Playwright tests (SKILL.md). 4. Sanitization: No automated sanitization is described; the skill relies on the 'reconciliation sheet' reviewed by a human operator.
[COMMAND_EXECUTION]: The skill references the use of standard development tools for environment management and testing within the project scope. Evidence: Instructions mention using git rev-parse --show-toplevel to verify the worktree and running Playwright for visual verification (SKILL.md).

mockup-parity