The Agent Skills Directory

[PROMPT_INJECTION]: While the skill interacts with untrusted external web content, it includes defensive instructions for the agent to "Trust nothing" and verify every finding with reproducible evidence, mitigating potential indirect prompt injection risks.- [COMMAND_EXECUTION]: The skill utilizes several CLI-based security tools for vulnerability scanning. It demonstrates security best practices by warning against the use of unverified community templates and recommending version pinning to mitigate risks associated with tool-specific vulnerabilities like CVE-2024-43405.

probe