The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides instructions to execute system-level commands and security tools including oscap, lynis, and inspec to perform compliance audits.
[PRIVILEGE_ESCALATION]: The identified audit tools and package management commands (apt install) require administrative or root privileges to access system configurations, logs, and installation directories.
[INDIRECT_PROMPT_INJECTION]: The skill involves reading and analyzing output files from external scanners (e.g., results.xml, lynis-report.dat). If these systems contain malicious data designed to look like scan findings, they could potentially influence the agent's remediation analysis.
Ingestion points: results.xml, report.html, lynis-report.dat, and standard output from security tools.
Boundary markers: No explicit boundary markers or instruction-ignoring delimiters are used when processing scanner output files.
Capability inventory: The skill has capabilities to install software (apt), run containers (docker), and execute shell commands for auditing.
Sanitization: There is no evidence of specific sanitization or validation of the report content before it is processed by the agent.

cis-benchmarks