The Agent Skills Directory

[SAFE]: The skill contains legitimate instructions for accessibility auditing and inclusive design verification without any malicious patterns.
[COMMAND_EXECUTION]: The documentation mentions the use of standard accessibility testing tools such as axe-core and Lighthouse via npx. These are referenced as part of a recommended manual workflow for the agent to follow when performing audits.
[REMOTE_CODE_EXECUTION]: The execution script scripts/run.sh is a standard wrapper that invokes a local Python runner. It does not perform any unauthorized remote code execution or network downloads.
[INDIRECT_PROMPT_INJECTION]: As an auditor, the skill is designed to process untrusted external data (web pages and code). This represents an attack surface for indirect prompt injection; however, the skill provides rigid reporting structures and specific criteria (WCAG 2.2) which helps guide the agent towards objective analysis rather than instruction following from the audited data.

testing-accessibility-auditor