accessibility-wcag
Warn
Audited by Gen Agent Trust Hub on Aug 10, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to autonomously run a shell command (
ls) targeting multiple hidden directories in the user's home folder and project root (e.g.,~/.claude/skills/,~/.agent/skills/,.agents/skills/). This is used to perform a 'companion check' to programmatically identify installed tools. - [PROMPT_INJECTION]: The instructions mandate that the agent must override its default response behavior by always starting the first response with a specific emoji ('🧢') and performing a conditional environment check upon activation.
- [EXTERNAL_DOWNLOADS]: The documentation and agent instructions recommend the installation of additional skills using 'npx skills add AbsolutelySkilled/AbsolutelySkilled', which involves fetching and executing code from an external repository.
Audit Metadata