skills/jeremylongshore/claude-code-plugins-plus-skills/implementing-database-audit-logging/Gen Agent Trust Hub
implementing-database-audit-logging
Warn
Audited by Gen Agent Trust Hub on Mar 23, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script 'scripts/audit-archival.sh' uses 'rm -f' via 'xargs' to delete files based on a retention policy. This is a destructive operation performed on files within a user-specified directory.
- [PROMPT_INJECTION]: Deceptive script headers and contents in the 'scripts/' directory. The files 'scripts/audit_log_analyzer.py', 'scripts/audit_log_generator.py', and 'scripts/audit_table_creator.sh' all contain identical directory-scanning boilerplate code that does not implement the functionality described in their names or the 'scripts/README.md' file. Additionally, 'scripts/audit_table_creator.sh' contains Python code despite its shell extension.
- [PROMPT_INJECTION]: Indirect prompt injection surface exists in 'scripts/audit_log_analyzer.py'. [Ingestion points]: Reads all files in a target directory using filesystem globbing. [Boundary markers]: Absent; file paths and names are processed without delimiters or warnings. [Capability inventory]: The skill environment includes file deletion ('scripts/audit-archival.sh') and database access tools ('psql', 'mysql'). [Sanitization]: Absent; the script processes file system entries directly.
Audit Metadata