The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from system logs which creates a surface for indirect prompt injection if logs contain attacker-controlled strings. 1. Ingestion points: Log files read from /var/log/ in scripts/agent.py and scripts/process.py. 2. Boundary markers: No delimiters or explicit instructions to ignore embedded content are provided. 3. Capability inventory: The skill has file-read and local file-write capabilities. 4. Sanitization: Log data is extracted using regular expressions without additional sanitization.
[COMMAND_EXECUTION]: The script agent.py contains references to destructive commands such as rm -rf / and dd. These are used as detection signatures for scanning command history files and are not executed by the agent.
[EXTERNAL_DOWNLOADS]: The skill references book.hacktricks.xyz for forensic methodology. This is a well-known cybersecurity knowledge base and its inclusion is contextually appropriate for the forensic analysis task.

performing-linux-log-forensics-investigation