The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from targets via the $ARGUMENTS variable across all phases of the security assessment and remediation process.
The data is interpolated into subagent prompts without the use of boundary markers (e.g., XML tags) or specific instructions to ignore malicious content embedded within the target code.
The skill enables high-impact capabilities, including automated code patching, infrastructure configuration changes, and the execution of penetration testing tools.
There is no mention of sanitizing or filtering input from external sources before processing.
[COMMAND_EXECUTION]: The skill utilizes the Task tool to execute a variety of security scanners and diagnostic utilities.
Tools listed include Semgrep, OWASP ZAP, Metasploit, and Burp Suite.
The instructions specifically authorize the agent to perform authenticated and unauthenticated penetration testing using custom exploits.
[DYNAMIC_EXECUTION]: The workflow involves the agent generating and applying security patches to remediate vulnerabilities.
Phase 2 instructions include fixing SQL injections, XSS, and authentication bypasses by modifying the backend and frontend code.
While this is the intended function of the skill, it involves the runtime generation and application of executable code based on analyzed findings.

security-scanning-security-hardening