The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze potentially untrusted data in the form of rendered HTML, interactive components, or design-system patterns to check for accessibility compliance.
Ingestion points: The skill instructions in SKILL.md direct the agent to review rendered HTML and interactive components.
Boundary markers: The instructions do not specify the use of delimiters or clear separation between the agent's instructions and the untrusted HTML data being analyzed.
Capability inventory: The skill appears to be limited to analysis and reporting; no high-risk capabilities such as network requests, file system writes, or subprocess execution are identified in the provided files.
Sanitization: There are no instructions for sanitizing or escaping the input HTML before it is processed by the agent.

button-name