The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/verify.sh executes arbitrary shell commands provided through the --command argument using bash -c. While this is the intended functionality for verifying command outcomes, it allows for the execution of any system command.
[DATA_EXFILTRATION]: The skill captures sensitive system metadata, including the full process list (ps aux) and network connection states (netstat or ss). This information is stored in a temporary state directory and included in reports, potentially exposing internal system configurations in logs.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes output from executed commands and file contents without sanitization. Ingestion points: scripts/verify.sh reads command output from command_output.txt and file content via grep. Boundary markers: None are implemented to isolate external content from agent instructions. Capability inventory: The skill can execute arbitrary shell commands, access the filesystem, and monitor system processes. Sanitization: No validation or escaping is applied to the ingested data.

testing-expected-results