The Agent Skills Directory

[METADATA_POISONING]: The skill metadata includes a verification: security_reviewed field. Users should be aware that this is a self-reported claim by the author and not an independent safety guarantee.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates browser automation and processes content from web applications, which constitutes an attack surface for indirect prompt injection.
Ingestion points: The skill captures DOM elements, web page content, and screenshots from the applications being tested.
Boundary markers: The provided documentation does not specify the use of delimiters or instructions to ignore embedded instructions within the processed web data.
Capability inventory: The tool has capabilities for file system writes (baseline and diff snapshots), network access (CI integration and Slack notifications), and full browser interaction.
Sanitization: No explicit sanitization or filtering of the content retrieved from external web pages is mentioned in the skill definition.
[COMMAND_EXECUTION]: The skill provides instructions for environment setup using npx and a vendor-specific CLI tool (clawhub). These commands are intended to install the skill and its dependencies from the vendor's infrastructure.

Cypress Visual Regression Testing Suite