skills/jeremylongshore/claude-code-plugins-plus-skills/comparing-database-schemas/Gen Agent Trust Hub
comparing-database-schemas
Warn
Audited by Gen Agent Trust Hub on Mar 12, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The script
scripts/generate_migration_script.pygenerates bash scripts from input content and programmatically grants them executable permissions. - Evidence: The
generate_scriptmethod inscripts/generate_migration_script.pywrites a shell script and callsfile_path.chmod(0o755). - [PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it processes untrusted database schemas and configuration files to generate executable scripts.
- Ingestion points: External content passed to
args.contentandargs.configinscripts/generate_migration_script.py, and files scanned inscripts/rollback_migration.py. - Boundary markers: The skill instructions do not specify any delimiters or safety warnings for handling embedded instructions in the processed data.
- Capability inventory: The skill is authorized to use
Bash(psql:*),Bash(mysql:*), andBash(mongosh:*)tools, along with file system write and chmod capabilities. - Sanitization: No sanitization or validation logic is present to filter malicious commands or instructions from the source data before script generation.
Audit Metadata