The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/check_accessibility.py uses subprocess.run to invoke the npx axe command for auditing websites. This involves executing system commands based on user-provided URLs, which is a core capability of the skill.
[EXTERNAL_DOWNLOADS]: The install.sh script installs several dependencies from well-known package registries including APT, NPM, and PyPI. This includes standard tools such as nodejs, python3, axe-core, lighthouse, and specific Python accessibility libraries. These are recognized services and their use for acquiring legitimate tools is considered standard practice.
[PROMPT_INJECTION]: The skill is designed to analyze external websites and files for accessibility compliance, creating a surface for indirect prompt injection.
Ingestion points: External web content from URLs or local HTML files provided by the user for auditing, processed via scripts/check_accessibility.py.
Boundary markers: The implementation does not include explicit markers or instructions to isolate the audited content from the agent's operating instructions.
Capability inventory: The skill utilizes subprocess.run to execute auditing commands and relies on network requests made by the underlying tools to fetch remote content.
Sanitization: No specific content filtering or sanitization is implemented to prevent malicious instructions embedded in the audited content from influencing the agent's behavior.

accessibility-compliance