The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit external code and web content provided via file paths or URLs, which are untrusted ingestion points.
Ingestion points: The skill accepts user-supplied files or URLs as arguments for accessibility auditing.
Boundary markers: The instructions do not include specific delimiters or instructions to ignore embedded prompts within the audited data.
Capability inventory: The skill uses Read, Grep, Glob, and Bash to analyze source files and execute auditing utilities.
Sanitization: No explicit sanitization or filtering of the ingested content is performed before analysis.
[EXTERNAL_DOWNLOADS]: The skill references and provides instructions for using standard, well-known accessibility auditing tools and repositories.
It suggests running axe-core (via axe-core-gems), Lighthouse, and Pa11y to perform automated checks.
It includes a reference to the official axe-core-gems repository on GitHub (github.com/dequelabs/axe-core-gems).
[COMMAND_EXECUTION]: The skill provides bash command snippets for the agent to run local auditing tools.
These commands include bundle exec rspec for running system specs and npx for executing accessibility CLI tools like lighthouse and pa11y.

accessibility-review