The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill instructs the agent to use a Node.js one-liner (node -e) to parse local Lighthouse audit reports (report.json). This is a legitimate technique to filter large datasets before processing them within the agent's context.
[COMMAND_EXECUTION]: Provides shell-based commands for parsing JSON files and interacting with local files, which are standard for developer auditing workflows.
[INDIRECT_PROMPT_INJECTION]: The skill processes content from external web pages and browser console logs to identify accessibility issues. This represents a potential surface for indirect prompt injection if a target page contains malicious instructions.
Ingestion points: Data enters the context via take_snapshot, list_console_messages, and lighthouse_audit results.
Boundary markers: The skill does not explicitly use boundary markers (like XML tags or delimiters) when ingesting this data.
Capability inventory: The skill possesses capabilities to execute browser scripts (evaluate_script), perform keyboard interactions (press_key), and run shell commands (node).
Sanitization: No explicit sanitization or filtering of the ingested external content is described before it is processed by the agent.

a11y-debugging