The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute the astrofy check --category accessibility command to perform automated audits. This is a primary function of the skill and appears legitimate within the context of a developer tool.
[EXTERNAL_DOWNLOADS]: Technical documentation within the skill references external resources from the W3C (World Wide Web Consortium) regarding Reflow, Focus Not Obscured, and Menus tutorials. These are trusted, well-known services for web accessibility standards.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted project data which creates a potential vector for indirect instructions.
Ingestion points: The skill reads local site routes, accessible trees, styles, and interactive flows, as well as project metadata from package.json, node_modules, and lockfiles.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore or isolate potentially malicious instructions embedded within the analyzed project files.
Capability inventory: The agent has the capability to execute shell commands (astrofy check) and read arbitrary files within the project directory.
Sanitization: No evidence of sanitization, filtering, or validation is present for the content extracted from the analyzed files.

astrofy-accessibility