The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to use shell commands like find and grep to analyze source code for accessibility violations. It also uses the open command to launch the official Accessibility Inspector tool. All commands are scoped to the local environment and are appropriate for the tool's auditing purpose.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted Swift source code from a user-provided directory, which acts as an ingestion point. Although the skill does not explicitly define sanitization or boundary markers for the audited content, the risk of the agent being misled by instructions hidden in the code is low and inherent to the task of code analysis. 1. Ingestion points: Swift source files within a local target path. 2. Boundary markers: None specified for the content of audited files. 3. Capability inventory: Access to local file search and application launching tools. 4. Sanitization: No sanitization of the content extracted from audited files.
[SAFE]: All external references and documentation links point to trusted organizations such as Apple and the W3C. The skill does not perform any remote code downloads, data exfiltration, or persistence-related actions.

auditing-accessibility