The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit external web pages, creating a surface for indirect prompt injection where an attacker could place malicious instructions in the text content of a page to influence the agent's behavior during the audit.
Ingestion points: contrast.js (lines 100-101) extracts value, placeholder, and textContent from DOM elements.
Boundary markers: None. The extracted text is directly included in the findings report without delimiters or instructions to ignore embedded content.
Capability inventory: The agent has the capability to execute JavaScript in a browser context and generate reports based on the findings.
Sanitization: The script performs basic whitespace normalization and truncates text to 40 characters, but it does not filter or sanitize the content for potential injection attacks.
[COMMAND_EXECUTION]: The skill explicitly instructs the agent to use a browser's evaluation capability to execute the provided contrast.js script. While this is the intended functionality for a web auditing tool, it involves executing code within a user's browser session.

color