The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from local project files (HTML, JSX, TSX, Vue, CSS) to perform accessibility audits. This ingestion path represents a potential surface for indirect prompt injection if malicious instructions are embedded in the project source code.
Ingestion points: Uses Grep, Glob, and Read tools to inspect files in the src/ directory.
Boundary markers: The instructions lack explicit delimiters or warnings to treat file content as untrusted data.
Capability inventory: The agent has access to powerful tools including Bash for command execution and Edit/Write for file system modifications.
Sanitization: There is no evidence of sanitization or filtering of the content retrieved from source files.
[COMMAND_EXECUTION]: The skill executes shell commands via Bash to detect the presence of accessibility tools (axe-core, pa11y, lighthouse, eslint plugins) and to perform diagnostic searches through the codebase using grep. It also instructs the agent to run automated audit tools like lighthouse and pa11y against a local development server.

accessibility-expert