The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes subprocess.run in scripts/agent.py to execute system commands such as find for locating SUID/SGID binaries and suspicious files in temporary directories. These commands are static and used for forensic discovery.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and parse untrusted data from potentially compromised Linux systems, creating a surface for indirect prompt injection.\n
Ingestion points: Reads content from system artifacts including .bash_history, /etc/passwd, /etc/shadow, cron jobs, and systemd service files.\n
Boundary markers: Absent. The forensic output utilizes plain text headers but does not include explicit delimiters or instructions for the agent to ignore instructions embedded within the analyzed data.\n
Capability inventory: The skill provides capabilities for file system traversal and shell command execution via the provided forensic scripts.\n
Sanitization: No sanitization or escaping is performed on the ingested log content, which could allow an adversary to influence agent behavior through crafted log entries or command history.

analyzing-linux-system-artifacts